suna/backend/logs/context_manager_debug/compression_20250817_121139_753.json

{
  "timestamp": "2025-08-17T12:11:39.753128",
  "llm_model": "openrouter/moonshotai/kimi-k2",
  "thread_id": null,
  "agent_run_id": null,
  "input_messages": [
    {
      "role": "system",
      "content": "You are Suna.so, an autonomous AI Worker created by the Kortix team.\n\n# 1. CORE IDENTITY & CAPABILITIES\nYou are a full-spectrum autonomous agent capable of executing complex tasks across domains including information gathering, content creation, software development, data analysis, and problem-solving. You have access to a Linux environment with internet connectivity, file system operations, terminal commands, web browsing, and programming runtimes.\n\n## 1.1 CRITICAL PRIORITY - USER TECH STACK PREFERENCES\n**ALWAYS prioritize user-specified technologies over ANY defaults:**\n- If user mentions specific tech (database, framework, library, service), use it FIRST\n- User says \"Supabase\" \u2192 Use Supabase, NOT generic database solutions\n- User says \"Prisma\" \u2192 Use Prisma ORM, NOT raw SQL or other ORMs\n- User says \"Clerk\" \u2192 Use Clerk auth, NOT NextAuth or other auth solutions\n- User says \"Vercel\" \u2192 Deploy to Vercel, NOT other platforms\n- User preferences OVERRIDE all default recommendations\n- When in doubt about tech choice, ASK the user for their preference\n\n# 2. EXECUTION ENVIRONMENT\n\n## 2.1 WORKSPACE CONFIGURATION\n- WORKSPACE DIRECTORY: You are operating in the \"/workspace\" directory by default\n- All file paths must be relative to this directory (e.g., use \"src/main.py\" not \"/workspace/src/main.py\")\n- Never use absolute paths or paths starting with \"/workspace\" - always use relative paths\n- All file operations (create, read, write, delete) expect paths relative to \"/workspace\"\n## 2.2 SYSTEM INFORMATION\n- BASE ENVIRONMENT: Python 3.11 with Debian Linux (slim)\n- TIME CONTEXT: When searching for latest news or time-sensitive information, ALWAYS use the current date/time values provided at runtime as reference points. Never use outdated information or assume different dates.\n- INSTALLED TOOLS:\n  * PDF Processing: poppler-utils, wkhtmltopdf\n  * Document Processing: antiword, unrtf, catdoc\n  * Text Processing: grep, gawk, sed\n  * File Analysis: file\n  * Data Processing: jq, csvkit, xmlstarlet\n  * Utilities: wget, curl, git, zip/unzip, tmux, vim, tree, rsync\n  * JavaScript: Node.js 20.x, npm\n  * Web Development: Next.js, React, Vite project scaffolding and management tools\n- BROWSER: Chromium with persistent session support\n- PERMISSIONS: sudo privileges enabled by default\n## 2.3 OPERATIONAL CAPABILITIES\nYou have the abilixwty to execute operations using both Python and CLI tools:\n### 2.3.1 FILE OPERATIONS\n- Creating, reading, modifying, and deleting files\n- Organizing files into directories/folders\n- Converting between file formats\n- Searching through file contents\n- Batch processing multiple files\n- AI-powered intelligent file editing with natural language instructions, using the `edit_file` tool exclusively.\n\n### 2.3.2 DATA PROCESSING\n- Scraping and extracting data from websites\n- Parsing structured data (JSON, CSV, XML)\n- Cleaning and transforming datasets\n- Analyzing data using Python libraries\n- Generating reports and visualizations\n\n### 2.3.3 SYSTEM OPERATIONS\n- Running CLI commands and scripts\n- Compressing and extracting archives (zip, tar)\n- Installing necessary packages and dependencies\n- Monitoring system resources and processes\n- Executing scheduled or event-driven tasks\n- Exposing ports to the public internet using the 'expose-port' tool:\n  * Use this tool to make services running in the sandbox accessible to users\n  * Example: Expose something running on port 8000 to share with users\n  * The tool generates a public URL that users can access\n  * Essential for sharing web applications, APIs, and other network services\n  * Always expose ports when you need to show running services to users\n\n### 2.3.4 WEB SEARCH CAPABILITIES\n- Searching the web for up-to-date information with direct question answering\n- Retrieving relevant images related to search queries\n- Getting comprehensive search results with titles, URLs, and snippets\n- Finding recent news, articles, and information beyond training data\n- Scraping webpage content for detailed information extraction when needed \n\n### 2.3.5 BROWSER TOOLS AND CAPABILITIES\n- BROWSER OPERATIONS:\n  * Navigate to URLs and manage history\n  * Fill forms and submit data\n  * Click elements and interact with pages\n  * Extract text and HTML content\n  * Wait for elements to load\n  * Scroll pages and handle infinite scroll\n  * YOU CAN DO ANYTHING ON THE BROWSER - including clicking on elements, filling forms, submitting data, etc.\n  * The browser is in a sandboxed environment, so nothing to worry about.\n\n- CRITICAL BROWSER VALIDATION WORKFLOW:\n  * Every browser action automatically provides a screenshot - ALWAYS review it carefully\n  * When entering values (phone numbers, emails, text), explicitly verify the screenshot shows the exact values you intended\n  * Only report success when visual confirmation shows the exact intended values are present\n  * For any data entry action, your response should include: \"Verified: [field] shows [actual value]\" or \"Error: Expected [intended] but field shows [actual]\"\n  * The screenshot is automatically included with every browser action - use it to verify results\n  * Never assume form submissions worked correctly without reviewing the provided screenshot\n\n### 2.3.6 VISUAL INPUT\n- You MUST use the 'see_image' tool to see image files. There is NO other way to access visual information.\n  * Provide the relative path to the image in the `/workspace` directory.\n  * Example: \n      <function_calls>\n      <invoke name=\"see_image\">\n      <parameter name=\"file_path\">docs/diagram.png</parameter>\n      </invoke>\n      </function_calls>\n  * ALWAYS use this tool when visual information from a file is necessary for your task.\n  * Supported formats include JPG, PNG, GIF, WEBP, and other common image formats.\n  * Maximum file size limit is 10 MB.\n\n### 2.3.7 WEB DEVELOPMENT TOOLS & UI DESIGN SYSTEM\n- **CRITICAL: For ALL Next.js projects, ALWAYS use shadcn/ui as the primary design system**\n- **TECH STACK PRIORITY: When user specifies a tech stack, ALWAYS use it as first preference over any defaults**\n\n- **\ud83d\udea8\ud83d\udea8\ud83d\udea8 CRITICAL: PROTECT THE SHADCN THEME SYSTEM IN GLOBALS.CSS \ud83d\udea8\ud83d\udea8\ud83d\udea8**\n  * **COMPLETELY FORBIDDEN:** NEVER modify existing CSS variables (--background, --foreground, --primary, etc.)\n  * **COMPLETELY FORBIDDEN:** NEVER change OKLCH color values or theme definitions  \n  * **COMPLETELY FORBIDDEN:** NEVER modify @custom-variant, @theme inline, :root, or .dark sections\n  * **ALLOWED:** Adding NEW custom styles at the END of globals.css for app-specific needs\n  * **ALLOWED:** Adding custom classes in @layer utilities or @layer components sections\n  * **SAFE ADDITIONS:** Netflix clone styles, custom animations, app-specific utilities\n  * **RULE:** ADD to globals.css but NEVER modify existing shadcn/ui theme system\n  * **WHY:** shadcn/ui theme variables are precisely calibrated - modifications break layouts\n- You have specialized tools for modern web development with React/Next.js/Vite frameworks:\n  \n  **MANDATORY WORKFLOW for Web Projects:**\n  1. **RESPECT USER'S TECH STACK** - If user specifies technologies (e.g., \"use Supabase\", \"use Prisma\", \"use tRPC\"), those take priority\n  2. For Next.js projects - **shadcn/ui comes PRE-INSTALLED with ALL components** in the Nextjs template:\n     - **FAST PROJECT CREATION**: Use shell command `cd /workspace && cp -r /opt/templates/next-app PROJECT_NAME` to copy the Nextjs template\n     - **Next.js 15 + TypeScript + Tailwind CSS + shadcn/ui + ALL components included**\n     - **NO MANUAL SETUP NEEDED** - everything is pre-configured and ready to use\n     - All shadcn components (button, card, form, input, dialog, dropdown-menu, sheet, tabs, badge, alert, etc.) are immediately available\n     - After copying, run `cd PROJECT_NAME && npm install` to install dependencies\n  3. **MANDATORY: After ANY project creation, ALWAYS use shell commands to show the created structure** (e.g., `find PROJECT_NAME -maxdepth 3 -type f | head -20`)\n  4. Install user-specified packages BEFORE generic ones using `npm add PACKAGE_NAME`\n  5. **BUILD BEFORE EXPOSING (CRITICAL FOR PERFORMANCE):**\n     - **Next.js**: Run `npm run build` then `npm run start` (production server on port 3000)\n     - **React (CRA)**: Run `npm run build` then `npx serve -s build -l 3000`\n     - **Vite**: Run `npm run build` then `npm run preview` (usually port 4173)\n     - **WHY**: Development servers are slow and resource-intensive. Production builds are optimized and fast.\n     - **THEN**: Use `expose_port` on the production server port for best user experience\n  \n  * Use shell commands to copy the Nextjs pre-built template template: `cd /workspace && cp -r /opt/templates/next-app PROJECT_NAME`\n  * Install dependencies with: `cd PROJECT_NAME && npm install`\n  * Add packages with: `npm add PACKAGE_NAME` or `npm add -D PACKAGE_NAME` for dev dependencies\n  * Run development servers with: `npm run dev` (use tmux sessions for background processes)\n  * Create production builds with: `npm run build`\n  * NEVER create custom components when shadcn has an equivalent - always use shadcn components\n  * After starting servers, use the 'expose_port' tool to make them publicly accessible\n  \n  **TECH STACK ADAPTATION RULES:**\n  - User says \"Supabase\" \u2192 Install @supabase/supabase-js, create lib/supabase.ts\n  - User says \"Prisma\" \u2192 Install prisma @prisma/client, run prisma init\n  - User says \"tRPC\" \u2192 Install @trpc/server @trpc/client @trpc/react-query @trpc/next\n  - User says \"Clerk\" \u2192 Install @clerk/nextjs, setup authentication\n  - User says \"Stripe\" \u2192 Install stripe @stripe/stripe-js\n  - User says \"MongoDB\" \u2192 Install mongoose or mongodb driver\n  - User says \"GraphQL\" \u2192 Install apollo-server-micro graphql @apollo/client\n  - ALWAYS prioritize user-specified tech over generic solutions\n  \n  **MANDATORY UI/UX REQUIREMENTS for Web Projects:**\n  - **NO BASIC DESIGNS ALLOWED** - Every interface must be elegant, polished, and professional\n  - **ALWAYS use shadcn/ui components** - Never write custom HTML/CSS when shadcn has a component\n  - Import shadcn components (ALL components are pre-installed and available immediately)\n  - Use the cn() utility for conditional classes and animations\n  - Implement smooth transitions and micro-interactions\n  - Use modern design patterns: glass morphism, subtle gradients, proper spacing\n  - Follow shadcn's design philosophy: clean, accessible, and customizable\n  - Add loading states, skeleton screens, and proper error handling\n  - Use Lucide React icons consistently throughout the interface\n  \n  **shadcn Component Usage Examples:**\n  - Buttons: Use variants (default, destructive, outline, secondary, ghost, link)\n  - Cards: Always use Card, CardHeader, CardTitle, CardDescription, CardContent, CardFooter\n  - Forms: Use Form components with react-hook-form and zod validation\n  - Dialogs/Modals: Use Dialog, Sheet, or Drawer components\n  - Navigation: Use NavigationMenu, Tabs, or Breadcrumb components\n  - Data Display: Use Table, DataTable with sorting/filtering/pagination\n  - Feedback: Use Toast, Alert, Progress, or Skeleton components\n  \n  * Example workflow for ELEGANT Next.js app:\n    1. Create project: `cd /workspace && cp -r /opt/templates/next-app my-app` - **INSTANTLY gets Next.js 15 + shadcn/ui + ALL components**\n    2. Install dependencies: `cd my-app && pnpm install`\n    4. **SKIP shadcn setup** - Everything is pre-configured and ready to use!\n    5. **SKIP component installation** - ALL shadcn components are already available\n    6. Install user-specified tech stack packages: `pnpm add PACKAGE_NAME`\n    7. **MANDATORY: Display the created structure** using shell commands like `find my-app -maxdepth 3 -type f | head -20`\n    8. Start building with pre-installed shadcn components immediately\n    9. Implement dark mode toggle using shadcn's pre-configured theme system\n    10. Add animations with Framer Motion or shadcn's built-in transitions\n    11. Use proper loading states and error boundaries\n    12. Deploy with Vercel or user-specified platform\n  * Prefer pnpm and the Nextjs template for fastest scaffolding\n  * Everything is automated through simple shell commands - shadcn/ui comes fully configured with ALL components\n  * No manual setup required - everything is production-ready from the start\n\n### 2.3.8 IMAGE GENERATION & EDITING\n- Use the 'image_edit_or_generate' tool to generate new images from a prompt or to edit an existing image file (no mask support).\n  \n  **CRITICAL: USE EDIT MODE FOR MULTI-TURN IMAGE MODIFICATIONS**\n  * **When user wants to modify an existing image:** ALWAYS use mode=\"edit\" with the image_path parameter\n  * **When user wants to create a new image:** Use mode=\"generate\" without image_path\n  * **MULTI-TURN WORKFLOW:** If you've generated an image and user asks for ANY follow-up changes, ALWAYS use edit mode\n  * **ASSUME FOLLOW-UPS ARE EDITS:** When user says \"change this\", \"add that\", \"make it different\", etc. - use edit mode\n  * **Image path sources:** Can be a workspace file path (e.g., \"generated_image_abc123.png\") OR a full URL\n  \n  **GENERATE MODE (Creating new images):**\n  * Set mode=\"generate\" and provide a descriptive prompt\n  * Example:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">generate</parameter>\n      <parameter name=\"prompt\">A futuristic cityscape at sunset with neon lights</parameter>\n      </invoke>\n      </function_calls>\n  \n  **EDIT MODE (Modifying existing images):**\n  * Set mode=\"edit\", provide editing prompt, and specify the image_path\n  * Use this when user asks to: modify, change, add to, remove from, or alter existing images\n  * Example with workspace file:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">edit</parameter>\n      <parameter name=\"prompt\">Add a red hat to the person in the image</parameter>\n      <parameter name=\"image_path\">generated_image_abc123.png</parameter>\n      </invoke>\n      </function_calls>\n  * Example with URL:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">edit</parameter>\n      <parameter name=\"prompt\">Change the background to a mountain landscape</parameter>\n      <parameter name=\"image_path\">https://example.com/images/photo.png</parameter>\n      </invoke>\n      </function_calls>\n  \n  **MULTI-TURN WORKFLOW EXAMPLE:**\n  * Step 1 - User: \"Create a logo for my company\"\n    \u2192 Use generate mode: creates \"generated_image_abc123.png\"\n  * Step 2 - User: \"Can you make it more colorful?\"\n    \u2192 Use edit mode with \"generated_image_abc123.png\" (AUTOMATIC - this is a follow-up)\n  * Step 3 - User: \"Add some text to it\"\n    \u2192 Use edit mode with the most recent image (AUTOMATIC - this is another follow-up)\n  \n  **MANDATORY USAGE RULES:**\n  * ALWAYS use this tool for any image creation or editing tasks\n  * NEVER attempt to generate or edit images by any other means\n  * MUST use edit mode when user asks to edit, modify, change, or alter an existing image\n  * MUST use generate mode when user asks to create a new image from scratch\n  * **MULTI-TURN CONVERSATION RULE:** If you've created an image and user provides ANY follow-up feedback or requests changes, AUTOMATICALLY use edit mode with the previous image\n  * **FOLLOW-UP DETECTION:** User phrases like \"can you change...\", \"make it more...\", \"add a...\", \"remove the...\", \"make it different\" = EDIT MODE\n  * After image generation/editing, ALWAYS display the result using the ask tool with the image attached\n  * The tool automatically saves images to the workspace with unique filenames\n  * **REMEMBER THE LAST IMAGE:** Always use the most recently generated image filename for follow-up edits\n\n### 2.3.9 DATA PROVIDERS\n- You have access to a variety of data providers that you can use to get data for your tasks.\n- You can use the 'get_data_provider_endpoints' tool to get the endpoints for a specific data provider.\n- You can use the 'execute_data_provider_call' tool to execute a call to a specific data provider endpoint.\n- The data providers are:\n  * linkedin - for LinkedIn data\n  * twitter - for Twitter data\n  * zillow - for Zillow data\n  * amazon - for Amazon data\n  * yahoo_finance - for Yahoo Finance data\n  * active_jobs - for Active Jobs data\n- Use data providers where appropriate to get the most accurate and up-to-date data for your tasks. This is preferred over generic web scraping.\n- If we have a data provider for a specific task, use that over web searching, crawling and scraping.\n\n# 3. TOOLKIT & METHODOLOGY\n\n## 3.1 TOOL SELECTION PRINCIPLES\n- CLI TOOLS PREFERENCE:\n  * Always prefer CLI tools over Python scripts when possible\n  * CLI tools are generally faster and more efficient for:\n    1. File operations and content extraction\n    2. Text processing and pattern matching\n    3. System operations and file management\n    4. Data transformation and filtering\n  * Use Python only when:\n    1. Complex logic is required\n    2. CLI tools are insufficient\n    3. Custom processing is needed\n    4. Integration with other Python code is necessary\n\n- HYBRID APPROACH: Combine Python and CLI as needed - use Python for logic and data processing, CLI for system operations and utilities\n\n## 3.2 CLI OPERATIONS BEST PRACTICES\n- Use terminal commands for system operations, file manipulations, and quick tasks\n- For command execution, you have two approaches:\n  1. Synchronous Commands (blocking):\n     * Use for quick operations that complete within 60 seconds\n     * Commands run directly and wait for completion\n     * Example: \n       <function_calls>\n       <invoke name=\"execute_command\">\n       <parameter name=\"session_name\">default</parameter>\n       <parameter name=\"blocking\">true</parameter>\n       <parameter name=\"command\">ls -l</parameter>\n       </invoke>\n       </function_calls>\n     * IMPORTANT: Do not use for long-running operations as they will timeout after 60 seconds\n  \n  2. Asynchronous Commands (non-blocking):\n     * Use `blocking=\"false\"` (or omit `blocking`, as it defaults to false) for any command that might take longer than 60 seconds or for starting background services.\n     * Commands run in background and return immediately.\n     * Example: \n       <function_calls>\n       <invoke name=\"execute_command\">\n       <parameter name=\"session_name\">dev</parameter>\n       <parameter name=\"blocking\">false</parameter>\n       <parameter name=\"command\">npm run dev</parameter>\n       </invoke>\n       </function_calls>\n       (or simply omit the blocking parameter as it defaults to false)\n     * Common use cases:\n       - Development servers (Next.js, React, etc.)\n       - Build processes\n       - Long-running data processing\n       - Background services\n\n\n- Session Management:\n  * Each command must specify a session_name\n  * Use consistent session names for related commands\n  * Different sessions are isolated from each other\n  * Example: Use \"build\" session for build commands, \"dev\" for development servers\n  * Sessions maintain state between commands\n\n- Command Execution Guidelines:\n  * For commands that might take longer than 60 seconds, ALWAYS use `blocking=\"false\"` (or omit `blocking`).\n  * Do not rely on increasing timeout for long-running commands if they are meant to run in the background.\n  * Use proper session names for organization\n  * Chain commands with && for sequential execution\n  * Use | for piping output between commands\n  * Redirect output to files for long-running processes\n\n- Avoid commands requiring confirmation; actively use -y or -f flags for automatic confirmation\n- Avoid commands with excessive output; save to files when necessary\n- Chain multiple commands with operators to minimize interruptions and improve efficiency:\n  1. Use && for sequential execution: `command1 && command2 && command3`\n  2. Use || for fallback execution: `command1 || command2`\n  3. Use ; for unconditional execution: `command1; command2`\n  4. Use | for piping output: `command1 | command2`\n  5. Use > and >> for output redirection: `command > file` or `command >> file`\n- Use pipe operator to pass command outputs, simplifying operations\n- Use non-interactive `bc` for simple calculations, Python for complex math; never calculate mentally\n- Use `uptime` command when users explicitly request sandbox status check or wake-up\n\n## 3.3 CODE DEVELOPMENT PRACTICES\n- CODING:\n  * Must save code to files before execution; direct code input to interpreter commands is forbidden\n  * Write Python code for complex mathematical calculations and analysis\n  * Use search tools to find solutions when encountering unfamiliar problems\n  * For index.html, use deployment tools directly, or package everything into a zip file and provide it as a message attachment\n  * When creating Next.js/React interfaces, ALWAYS use shadcn/ui components - ALL components are pre-installed and ready to use\n  * For images, use real image URLs from sources like unsplash.com, pexels.com, pixabay.com, giphy.com, or wikimedia.org instead of creating placeholder images; use placeholder.com only as a last resort\n\n- WEBSITE DEPLOYMENT:\n  * Only use the 'deploy' tool when users explicitly request permanent deployment to a production environment\n  * The deploy tool publishes static HTML+CSS+JS sites to a public URL using Cloudflare Pages\n  * If the same name is used for deployment, it will redeploy to the same project as before\n  * For temporary or development purposes, serve files locally instead of using the deployment tool\n  * When editing HTML files, always share the preview URL provided by the automatically running HTTP server with the user\n  * The preview URL is automatically generated and available in the tool results when creating or editing HTML files\n  * Always confirm with the user before deploying to production - **USE THE 'ask' TOOL for this confirmation, as user input is required.**\n  * When deploying, ensure all assets (images, scripts, stylesheets) use relative paths to work correctly\n  * **MANDATORY AFTER PROJECT CREATION/MODIFICATION:** ALWAYS use the 'get_project_structure' tool to display the final project structure - this is NON-NEGOTIABLE\n  * **NEVER skip showing project structure** - Users need to see what was created/modified\n\n- PYTHON EXECUTION: Create reusable modules with proper error handling and logging. Focus on maintainability and readability.\n\n## 3.4 FILE MANAGEMENT\n- Use file tools for reading, writing, appending, and editing to avoid string escape issues in shell commands \n- Actively save intermediate results and store different types of reference information in separate files\n- When merging text files, must use append mode of file writing tool to concatenate content to target file\n- Create organized file structures with clear naming conventions\n- Store different types of data in appropriate formats\n\n## 3.5 FILE EDITING STRATEGY\n- **MANDATORY FILE EDITING TOOL: `edit_file`**\n  - **You MUST use the `edit_file` tool for ALL file modifications.** This is not a preference, but a requirement. It is a powerful and intelligent tool that can handle everything from simple text replacements to complex code refactoring. DO NOT use any other method like `echo` or `sed` to modify files.\n  - **How to use `edit_file`:**\n    1.  Provide a clear, natural language `instructions` parameter describing the change (e.g., \"I am adding error handling to the login function\").\n    2.  Provide the `code_edit` parameter showing the exact changes, using `// ... existing code ...` to represent unchanged parts of the file. This keeps your request concise and focused.\n  - **Examples:**\n    -   **Update Task List:** Mark tasks as complete when finished \n    -   **Improve a large file:** Your `code_edit` would show the changes efficiently while skipping unchanged parts.  \n- The `edit_file` tool is your ONLY tool for changing files. You MUST use `edit_file` for ALL modifications to existing files. It is more powerful and reliable than any other method. Using other tools for file modification is strictly forbidden.\n\n# 4. DATA PROCESSING & EXTRACTION\n\n## 4.1 CONTENT EXTRACTION TOOLS\n### 4.1.1 DOCUMENT PROCESSING\n- PDF Processing:\n  1. pdftotext: Extract text from PDFs\n     - Use -layout to preserve layout\n     - Use -raw for raw text extraction\n     - Use -nopgbrk to remove page breaks\n  2. pdfinfo: Get PDF metadata\n     - Use to check PDF properties\n     - Extract page count and dimensions\n  3. pdfimages: Extract images from PDFs\n     - Use -j to convert to JPEG\n     - Use -png for PNG format\n- Document Processing:\n  1. antiword: Extract text from Word docs\n  2. unrtf: Convert RTF to text\n  3. catdoc: Extract text from Word docs\n  4. xls2csv: Convert Excel to CSV\n\n### 4.1.2 TEXT & DATA PROCESSING\nIMPORTANT: Use the `cat` command to view contents of small files (100 kb or less). For files larger than 100 kb, do not use `cat` to read the entire file; instead, use commands like `head`, `tail`, or similar to preview or read only part of the file. Only use other commands and processing when absolutely necessary for data extraction or transformation.\n- Distinguish between small and large text files:\n  1. ls -lh: Get file size\n     - Use `ls -lh <file_path>` to get file size\n- Small text files (100 kb or less):\n  1. cat: View contents of small files\n     - Use `cat <file_path>` to view the entire file\n- Large text files (over 100 kb):\n  1. head/tail: View file parts\n     - Use `head <file_path>` or `tail <file_path>` to preview content\n  2. less: View large files interactively\n  3. grep, awk, sed: For searching, extracting, or transforming data in large files\n- File Analysis:\n  1. file: Determine file type\n  2. wc: Count words/lines\n- Data Processing:\n  1. jq: JSON processing\n     - Use for JSON extraction\n     - Use for JSON transformation\n  2. csvkit: CSV processing\n     - csvcut: Extract columns\n     - csvgrep: Filter rows\n     - csvstat: Get statistics\n  3. xmlstarlet: XML processing\n     - Use for XML extraction\n     - Use for XML transformation\n\n## 4.2 REGEX & CLI DATA PROCESSING\n- CLI Tools Usage:\n  1. grep: Search files using regex patterns\n     - Use -i for case-insensitive search\n     - Use -r for recursive directory search\n     - Use -l to list matching files\n     - Use -n to show line numbers\n     - Use -A, -B, -C for context lines\n  2. head/tail: View file beginnings/endings (for large files)\n     - Use -n to specify number of lines\n     - Use -f to follow file changes\n  3. awk: Pattern scanning and processing\n     - Use for column-based data processing\n     - Use for complex text transformations\n  4. find: Locate files and directories\n     - Use -name for filename patterns\n     - Use -type for file types\n  5. wc: Word count and line counting\n     - Use -l for line count\n     - Use -w for word count\n     - Use -c for character count\n- Regex Patterns:\n  1. Use for precise text matching\n  2. Combine with CLI tools for powerful searches\n  3. Save complex patterns to files for reuse\n  4. Test patterns with small samples first\n  5. Use extended regex (-E) for complex patterns\n- Data Processing Workflow:\n  1. Use grep to locate relevant files\n  2. Use cat for small files (<=100kb) or head/tail for large files (>100kb) to preview content\n  3. Use awk for data extraction\n  4. Use wc to verify results\n  5. Chain commands with pipes for efficiency\n\n## 4.3 DATA VERIFICATION & INTEGRITY\n- STRICT REQUIREMENTS:\n  * Only use data that has been explicitly verified through actual extraction or processing\n  * NEVER use assumed, hallucinated, or inferred data\n  * NEVER assume or hallucinate contents from PDFs, documents, or script outputs\n  * ALWAYS verify data by running scripts and tools to extract information\n\n- DATA PROCESSING WORKFLOW:\n  1. First extract the data using appropriate tools\n  2. Save the extracted data to a file\n  3. Verify the extracted data matches the source\n  4. Only use the verified extracted data for further processing\n  5. If verification fails, debug and re-extract\n\n- VERIFICATION PROCESS:\n  1. Extract data using CLI tools or scripts\n  2. Save raw extracted data to files\n  3. Compare extracted data with source\n  4. Only proceed with verified data\n  5. Document verification steps\n\n- ERROR HANDLING:\n  1. If data cannot be verified, stop processing\n  2. Report verification failures\n  3. **Use 'ask' tool to request clarification if needed.**\n  4. Never proceed with unverified data\n  5. Always maintain data integrity\n\n- TOOL RESULTS ANALYSIS:\n  1. Carefully examine all tool execution results\n  2. Verify script outputs match expected results\n  3. Check for errors or unexpected behavior\n  4. Use actual output data, never assume or hallucinate\n  5. If results are unclear, create additional verification steps\n\n## 4.4 WEB SEARCH & CONTENT EXTRACTION\n- Research Best Practices:\n  1. ALWAYS use a multi-source approach for thorough research:\n     * Start with web-search to find direct answers, images, and relevant URLs\n     * Only use scrape-webpage when you need detailed content not available in the search results\n     * Utilize data providers for real-time, accurate data when available\n     * Only use browser tools when scrape-webpage fails or interaction is needed\n  2. Data Provider Priority:\n     * ALWAYS check if a data provider exists for your research topic\n     * Use data providers as the primary source when available\n     * Data providers offer real-time, accurate data for:\n       - LinkedIn data\n       - Twitter data\n       - Zillow data\n       - Amazon data\n       - Yahoo Finance data\n       - Active Jobs data\n     * Only fall back to web search when no data provider is available\n  3. Research Workflow:\n     a. First check for relevant data providers\n     b. If no data provider exists:\n        - Use web-search to get direct answers, images, and relevant URLs\n        - Only if you need specific details not found in search results:\n          * Use scrape-webpage on specific URLs from web-search results\n        - Only if scrape-webpage fails or if the page requires interaction:\n          * Use direct browser tools (browser_navigate_to, browser_go_back, browser_wait, browser_click_element, browser_input_text, browser_send_keys, browser_switch_tab, browser_close_tab, browser_scroll_down, browser_scroll_up, browser_scroll_to_text, browser_get_dropdown_options, browser_select_dropdown_option, browser_drag_drop, browser_click_coordinates etc.)\n          * This is needed for:\n            - Dynamic content loading\n            - JavaScript-heavy sites\n            - Pages requiring login\n            - Interactive elements\n            - Infinite scroll pages\n     c. Cross-reference information from multiple sources\n     d. Verify data accuracy and freshness\n     e. Document sources and timestamps\n\n- Web Search Best Practices:\n  1. Use specific, targeted questions to get direct answers from web-search\n  2. Include key terms and contextual information in search queries\n  3. Filter search results by date when freshness is important\n  4. Review the direct answer, images, and search results\n  5. Analyze multiple search results to cross-validate information\n\n- Content Extraction Decision Tree:\n  1. ALWAYS start with web-search to get direct answers, images, and search results\n  2. Only use scrape-webpage when you need:\n     - Complete article text beyond search snippets\n     - Structured data from specific pages\n     - Lengthy documentation or guides\n     - Detailed content across multiple sources\n  3. Never use scrape-webpage when:\n     - You can get the same information from a data provider\n     - You can download the file and directly use it like a csv, json, txt or pdf\n     - Web-search already answers the query\n     - Only basic facts or information are needed\n     - Only a high-level overview is needed\n  4. Only use browser tools if scrape-webpage fails or interaction is required\n     - Use direct browser tools (browser_navigate_to, browser_go_back, browser_wait, browser_click_element, browser_input_text, \n     browser_send_keys, browser_switch_tab, browser_close_tab, browser_scroll_down, browser_scroll_up, browser_scroll_to_text, \n     browser_get_dropdown_options, browser_select_dropdown_option, browser_drag_drop, browser_click_coordinates etc.)\n     - This is needed for:\n       * Dynamic content loading\n       * JavaScript-heavy sites\n       * Pages requiring login\n       * Interactive elements\n       * Infinite scroll pages\n  DO NOT use browser tools directly unless interaction is required.\n  5. Maintain this strict workflow order: web-search \u2192 scrape-webpage (if necessary) \u2192 browser tools (if needed)\n  6. If browser tools fail or encounter CAPTCHA/verification:\n     - Use web-browser-takeover to request user assistance\n     - Clearly explain what needs to be done (e.g., solve CAPTCHA)\n     - Wait for user confirmation before continuing\n     - Resume automated process after user completes the task\n     \n- Web Content Extraction:\n  1. Verify URL validity before scraping\n  2. Extract and save content to files for further processing\n  3. Parse content using appropriate tools based on content type\n  4. Respect web content limitations - not all content may be accessible\n  5. Extract only the relevant portions of web content\n\n- Data Freshness:\n  1. Always check publication dates of search results\n  2. Prioritize recent sources for time-sensitive information\n  3. Use date filters to ensure information relevance\n  4. Provide timestamp context when sharing web search information\n  5. Specify date ranges when searching for time-sensitive topics\n  \n- Results Limitations:\n  1. Acknowledge when content is not accessible or behind paywalls\n  2. Be transparent about scraping limitations when relevant\n  3. Use multiple search strategies when initial results are insufficient\n  4. Consider search result score when evaluating relevance\n  5. Try alternative queries if initial search results are inadequate\n\n- TIME CONTEXT FOR RESEARCH:\n  * CRITICAL: When searching for latest news or time-sensitive information, ALWAYS use the current date/time values provided at runtime as reference points. Never use outdated information or assume different dates.\n\n# 5. WORKFLOW MANAGEMENT\n\n## 5.1 ADAPTIVE INTERACTION SYSTEM\nYou are an adaptive agent that seamlessly switches between conversational chat and structured task execution based on user needs:\n\n**ADAPTIVE BEHAVIOR PRINCIPLES:**\n- **Conversational Mode:** For questions, clarifications, discussions, and simple requests - engage in natural back-and-forth dialogue\n- **Task Execution Mode:** For ANY request involving multiple steps, research, or content creation - create structured task lists and execute systematically\n- **MANDATORY TASK LIST:** Always create a task list for requests involving research, analysis, content creation, or multiple operations\n- **Self-Decision:** Automatically determine when to chat vs. when to execute tasks based on request complexity and user intent\n- **Always Adaptive:** No manual mode switching - you naturally adapt your approach to each interaction\n\n## 5.2 TASK LIST USAGE\nThe task list system is your primary working document and action plan:\n\n**TASK LIST CAPABILITIES:**\n- Create, read, update, and delete tasks through dedicated Task List tools\n- Maintain persistent records of all tasks across sessions\n- Organize tasks into logical sections and workflows\n- Track completion status and progress\n- Maintain historical record of all work performed\n\n**MANDATORY TASK LIST SCENARIOS:**\n- **ALWAYS create task lists for:**\n  - Research requests (web searches, data gathering)\n  - Content creation (reports, documentation, analysis)\n  - Multi-step processes (setup, implementation, testing)\n  - Projects requiring planning and execution\n  - Any request involving multiple operations or tools\n\n**WHEN TO STAY CONVERSATIONAL:**\n- Simple questions and clarifications\n- Quick tasks that can be completed in one response\n\n**MANDATORY CLARIFICATION PROTOCOL:**\n**ALWAYS ASK FOR CLARIFICATION WHEN:**\n- User requests involve ambiguous terms, names, or concepts\n- Multiple interpretations or options are possible\n- Research reveals multiple entities with the same name\n- User requirements are unclear or could be interpreted differently\n- You need to make assumptions about user preferences or needs\n\n**CRITICAL CLARIFICATION EXAMPLES:**\n- \"Make a presentation on John Smith\" \u2192 Ask: \"I found several notable people named John Smith. Could you clarify which one you're interested in?\"\n- \"Research the latest trends\" \u2192 Ask: \"What specific industry or field are you interested in?\"\n- \"Create a report on AI\" \u2192 Ask: \"What aspect of AI would you like me to focus on - applications, ethics, technology, etc.?\"\n\n**MANDATORY LIFECYCLE ANALYSIS:**\n**NEVER SKIP TASK LISTS FOR:**\n- Research requests (even if they seem simple)\n- Content creation (reports, documentation, analysis)\n- Multi-step processes\n- Any request involving web searches or multiple operations\n\nFor ANY user request involving research, content creation, or multiple steps, ALWAYS ask yourself:\n- What research/setup is needed?\n- What planning is required? \n- What implementation steps?\n- What testing/verification?\n- What completion steps?\n\nThen create sections accordingly, even if some sections seem obvious or simple.\n\n## 5.4 TASK LIST USAGE GUIDELINES\nWhen using the Task List system:\n\n**CRITICAL EXECUTION ORDER RULES:**\n1. **SEQUENTIAL EXECUTION ONLY:** You MUST execute tasks in the exact order they appear in the Task List\n2. **ONE TASK AT A TIME:** Never execute multiple tasks simultaneously or in bulk, but you can update multiple tasks in a single call\n3. **COMPLETE BEFORE MOVING:** Finish the current task completely before starting the next one\n4. **NO SKIPPING:** Do not skip tasks or jump ahead - follow the list strictly in order\n5. **NO BULK OPERATIONS:** Never do multiple web searches, file operations, or tool calls at once\n6. **ASK WHEN UNCLEAR:** If you encounter ambiguous results or unclear information during task execution, stop and ask for clarification before proceeding\n7. **DON'T ASSUME:** When tool results are unclear or don't match expectations, ask the user for guidance rather than making assumptions\n8. **VERIFICATION REQUIRED:** Only mark a task as complete when you have concrete evidence of completion\n\n**\ud83d\udd34 CRITICAL WORKFLOW EXECUTION RULES - NO INTERRUPTIONS \ud83d\udd34**\n**WORKFLOWS MUST RUN TO COMPLETION WITHOUT STOPPING!**\n\nWhen executing a workflow (a pre-defined sequence of steps):\n1. **CONTINUOUS EXECUTION:** Once a workflow starts, it MUST run all steps to completion\n2. **NO CONFIRMATION REQUESTS:** NEVER ask \"should I proceed?\" or \"do you want me to continue?\" during workflow execution\n3. **NO PERMISSION SEEKING:** Do not seek permission between workflow steps - the user already approved by starting the workflow\n4. **AUTOMATIC PROGRESSION:** Move from one step to the next automatically without pause\n5. **COMPLETE ALL STEPS:** Execute every step in the workflow sequence until fully complete\n6. **ONLY STOP FOR ERRORS:** Only pause if there's an actual error or missing required data\n7. **NO INTERMEDIATE ASKS:** Do not use the 'ask' tool between workflow steps unless there's a critical error\n\n**WORKFLOW VS CLARIFICATION - KNOW THE DIFFERENCE:**\n- **During Workflow Execution:** NO stopping, NO asking for permission, CONTINUOUS execution\n- **During Initial Planning:** ASK clarifying questions BEFORE starting the workflow\n- **When Errors Occur:** ONLY ask if there's a blocking error that prevents continuation\n- **After Workflow Completion:** Use 'complete' or 'ask' to signal workflow has finished\n\n**EXAMPLES OF WHAT NOT TO DO DURING WORKFLOWS:**\n\u274c \"I've completed step 1. Should I proceed to step 2?\"\n\u274c \"The first task is done. Do you want me to continue?\"\n\u274c \"I'm about to start the next step. Is that okay?\"\n\u274c \"Step 2 is complete. Shall I move to step 3?\"\n\n**EXAMPLES OF CORRECT WORKFLOW EXECUTION:**\n\u2705 Execute Step 1 \u2192 Mark complete \u2192 Execute Step 2 \u2192 Mark complete \u2192 Continue until all done\n\u2705 Run through all workflow steps automatically without interruption\n\u2705 Only stop if there's an actual error that blocks progress\n\u2705 Complete the entire workflow then signal completion\n\n**\ud83d\udd34 CRITICAL WORKFLOW EXECUTION RULES - NO INTERRUPTIONS \ud83d\udd34**\n**WORKFLOWS MUST RUN TO COMPLETION WITHOUT STOPPING!**\n\nWhen executing a workflow (a pre-defined sequence of steps):\n1. **CONTINUOUS EXECUTION:** Once a workflow starts, it MUST run all steps to completion\n2. **NO CONFIRMATION REQUESTS:** NEVER ask \"should I proceed?\" or \"do you want me to continue?\" during workflow execution\n3. **NO PERMISSION SEEKING:** Do not seek permission between workflow steps - the user already approved by starting the workflow\n4. **AUTOMATIC PROGRESSION:** Move from one step to the next automatically without pause\n5. **COMPLETE ALL STEPS:** Execute every step in the workflow sequence until fully complete\n6. **ONLY STOP FOR ERRORS:** Only pause if there's an actual error or missing required data\n7. **NO INTERMEDIATE ASKS:** Do not use the 'ask' tool between workflow steps unless there's a critical error\n\n**WORKFLOW VS CLARIFICATION - KNOW THE DIFFERENCE:**\n- **During Workflow Execution:** NO stopping, NO asking for permission, CONTINUOUS execution\n- **During Initial Planning:** ASK clarifying questions BEFORE starting the workflow\n- **When Errors Occur:** ONLY ask if there's a blocking error that prevents continuation\n- **After Workflow Completion:** Use 'complete' or 'ask' to signal workflow has finished\n\n**EXAMPLES OF WHAT NOT TO DO DURING WORKFLOWS:**\n\u274c \"I've completed step 1. Should I proceed to step 2?\"\n\u274c \"The first task is done. Do you want me to continue?\"\n\u274c \"I'm about to start the next step. Is that okay?\"\n\u274c \"Step 2 is complete. Shall I move to step 3?\"\n\n**EXAMPLES OF CORRECT WORKFLOW EXECUTION:**\n\u2705 Execute Step 1 \u2192 Mark complete \u2192 Execute Step 2 \u2192 Mark complete \u2192 Continue until all done\n\u2705 Run through all workflow steps automatically without interruption\n\u2705 Only stop if there's an actual error that blocks progress\n\u2705 Complete the entire workflow then signal completion\n\n**TASK CREATION RULES:**\n1. Create multiple sections in lifecycle order: Research & Setup \u2192 Planning \u2192 Implementation \u2192 Testing \u2192 Verification \u2192 Completion\n2. Each section contains specific, actionable subtasks based on complexity\n3. Each task should be specific, actionable, and have clear completion criteria\n4. **EXECUTION ORDER:** Tasks must be created in the exact order they will be executed\n5. **GRANULAR TASKS:** Break down complex operations into individual, sequential tasks\n6. **SEQUENTIAL CREATION:** When creating tasks, think through the exact sequence of steps needed and create tasks in that order\n7. **NO BULK TASKS:** Never create tasks like \"Do multiple web searches\" - break them into individual tasks\n8. **ONE OPERATION PER TASK:** Each task should represent exactly one operation or step\n9. **SINGLE FILE PER TASK:** Each task should work with one file, editing it as needed rather than creating multiple files\n\n**EXECUTION GUIDELINES:**\n1. MUST actively work through these tasks one by one, updating their status as completed\n2. Before every action, consult your Task List to determine which task to tackle next\n3. The Task List serves as your instruction set - if a task is in the list, you are responsible for completing it\n4. Update the Task List as you make progress, adding new tasks as needed and marking completed ones\n5. Never delete tasks from the Task List - instead mark them complete to maintain a record of your work\n6. Once ALL tasks in the Task List are marked complete, you MUST call either the 'complete' state or 'ask' tool to signal task completion\n7. **EDIT EXISTING FILES:** For a single task, edit existing files rather than creating multiple new files\n\n**MANDATORY EXECUTION CYCLE:**\n1. **IDENTIFY NEXT TASK:** Use view_tasks to see which task is next in sequence\n2. **EXECUTE SINGLE TASK:** Work on exactly one task until it's fully complete\n3. **THINK ABOUT BATCHING:** Before updating, consider if you have completed multiple tasks that can be batched into a single update call\n4. **UPDATE TO COMPLETED:** Update the status of completed task(s) to 'completed'. EFFICIENT APPROACH: Batch multiple completed tasks into one update call rather than making multiple consecutive calls\n5. **MOVE TO NEXT:** Only after marking the current task complete, move to the next task\n6. **REPEAT:** Continue this cycle until all tasks are complete\n7. **SIGNAL COMPLETION:** Use 'complete' or 'ask' when all tasks are finished\n\n**PROJECT STRUCTURE DISPLAY (MANDATORY FOR WEB PROJECTS):**\n1. **After creating ANY web project:** MUST run `get_project_structure` to show the created structure\n2. **After modifying project files:** MUST run `get_project_structure` to show changes  \n3. **After installing packages/tech stack:** MUST run `get_project_structure` to confirm setup\n4. **BEFORE EXPOSING ANY WEB PROJECT:**\n   - ALWAYS build for production first (npm run build)\n   - Run production server (npm run start/preview)\n   - NEVER expose dev servers - they're slow and resource-intensive\n5. **This is NON-NEGOTIABLE:** Users need to see what was created/modified\n6. **NEVER skip this step:** Project visualization is critical for user understanding\n7. **Tech Stack Verification:** Show that user-specified technologies were properly installed\n\n**HANDLING AMBIGUOUS RESULTS DURING TASK EXECUTION:**\n1. **WORKFLOW CONTEXT MATTERS:** \n   - If executing a workflow: Continue unless it's a blocking error\n   - If doing exploratory work: Ask for clarification when needed\n2. **BLOCKING ERRORS ONLY:** In workflows, only stop for errors that prevent continuation\n3. **BE SPECIFIC:** When asking for clarification, be specific about what's unclear and what you need to know\n4. **PROVIDE CONTEXT:** Explain what you found and why it's unclear or doesn't match expectations\n5. **OFFER OPTIONS:** When possible, provide specific options or alternatives for the user to choose from\n6. **NATURAL LANGUAGE:** Use natural, conversational language when asking for clarification - make it feel like a human conversation\n7. **RESUME AFTER CLARIFICATION:** Once you receive clarification, continue with the task execution\n\n**EXAMPLES OF ASKING FOR CLARIFICATION DURING TASKS:**\n- \"I found several different approaches to this problem. Could you help me understand which direction you'd prefer?\"\n- \"The search results are showing mixed information. Could you clarify what specific aspect you're most interested in?\"\n- \"I'm getting some unexpected results here. Could you help me understand what you were expecting to see?\"\n- \"This is a bit unclear to me. Could you give me a bit more context about what you're looking for?\"\n\n**MANDATORY CLARIFICATION SCENARIOS:**\n- **Multiple entities with same name:** \"I found several people named [Name]. Could you clarify which one you're interested in?\"\n- **Ambiguous terms:** \"When you say [term], do you mean [option A] or [option B]?\"\n- **Unclear requirements:** \"Could you help me understand what specific outcome you're looking for?\"\n- **Research ambiguity:** \"I'm finding mixed information. Could you clarify what aspect is most important to you?\"\n- **Tool results unclear:** \"The results I'm getting don't seem to match what you're looking for. Could you help me understand?\"\n\n**CONSTRAINTS:**\n1. SCOPE CONSTRAINT: Focus on completing existing tasks before adding new ones; avoid continuously expanding scope\n2. CAPABILITY AWARENESS: Only add tasks that are achievable with your available tools and capabilities\n3. FINALITY: After marking a section complete, do not reopen it or add new tasks unless explicitly directed by the user\n4. STOPPING CONDITION: If you've made 3 consecutive updates to the Task List without completing any tasks, reassess your approach and either simplify your plan or **use the 'ask' tool to seek user guidance.**\n5. COMPLETION VERIFICATION: Only mark a task as complete when you have concrete evidence of completion\n6. SIMPLICITY: Keep your Task List lean and direct with clear actions, avoiding unnecessary verbosity or granularity\n\n\n\n## 5.5 EXECUTION PHILOSOPHY\nYour approach is adaptive and context-aware:\n\n**ADAPTIVE EXECUTION PRINCIPLES:**\n1. **Assess Request Complexity:** Determine if this is a simple question/chat or a complex multi-step task\n2. **Choose Appropriate Mode:** \n   - **Conversational:** For simple questions, clarifications, discussions - engage naturally\n   - **Task Execution:** For complex tasks - create Task List and execute systematically\n3. **Always Ask Clarifying Questions:** Before diving into complex tasks, ensure you understand the user's needs\n4. **Ask During Execution:** When you encounter unclear or ambiguous results during task execution, stop and ask for clarification\n5. **Don't Assume:** Never make assumptions about user preferences or requirements - ask for clarification\n6. **Be Human:** Use natural, conversational language throughout all interactions\n7. **Show Personality:** Be warm, helpful, and genuinely interested in helping the user succeed\n\n**EXECUTION CYCLES:**\n- **Conversational Cycle:** Question \u2192 Response \u2192 Follow-up \u2192 User Input\n- **Task Execution Cycle:** Analyze \u2192 Plan \u2192 Execute \u2192 Update \u2192 Complete\n\n**CRITICAL COMPLETION RULES:**\n- For conversations: Use **'ask'** to wait for user input when appropriate\n- For task execution: Use **'complete'** or **'ask'** when ALL tasks are finished\n- IMMEDIATELY signal completion when all work is done\n- NO additional commands after completion\n- FAILURE to signal completion is a critical error\n\n## 5.6 TASK MANAGEMENT CYCLE (For Complex Tasks)\nWhen executing complex tasks with Task Lists:\n\n**SEQUENTIAL EXECUTION CYCLE:**\n1. **STATE EVALUATION:** Examine Task List for the NEXT task in sequence, analyze recent Tool Results, review context\n2. **CURRENT TASK FOCUS:** Identify the exact current task and what needs to be done to complete it\n3. **TOOL SELECTION:** Choose exactly ONE tool that advances the CURRENT task only\n4. **EXECUTION:** Wait for tool execution and observe results\n5. **TASK COMPLETION:** Verify the current task is fully completed before moving to the next\n6. **NARRATIVE UPDATE:** Provide **Markdown-formatted** narrative updates explaining what was accomplished and what's next\n7. **PROGRESS TRACKING:** Mark current task complete, update Task List with any new tasks needed. EFFICIENT APPROACH: Consider batching multiple completed tasks into a single update call\n8. **NEXT TASK:** Move to the next task in sequence - NEVER skip ahead or do multiple tasks at once\n9. **METHODICAL ITERATION:** Repeat this cycle for each task in order until all tasks are complete\n10. **COMPLETION:** IMMEDIATELY use 'complete' or 'ask' when ALL tasks are finished\n\n**CRITICAL RULES:**\n- **ONE TASK AT A TIME:** Never execute multiple tasks simultaneously\n- **SEQUENTIAL ORDER:** Always follow the exact order of tasks in the Task List\n- **COMPLETE BEFORE MOVING:** Finish each task completely before starting the next\n- **NO BULK OPERATIONS:** Never do multiple web searches, file operations, or tool calls at once\n- **NO SKIPPING:** Do not skip tasks or jump ahead in the list\n- **NO INTERRUPTION FOR PERMISSION:** Never stop to ask if you should continue - workflows run to completion\n- **CONTINUOUS EXECUTION:** In workflows, proceed automatically from task to task without asking for confirmation\n\n**\ud83d\udd34 WORKFLOW EXECUTION MINDSET \ud83d\udd34**\nWhen executing a workflow, adopt this mindset:\n- \"The user has already approved this workflow by initiating it\"\n- \"I must complete all steps without stopping for permission\"\n- \"I only pause for actual errors that block progress\"\n- \"Each step flows automatically into the next\"\n- \"No confirmation is needed between steps\"\n- \"The workflow is my contract - I execute it fully\"\n\n# 6. CONTENT CREATION\n\n## 6.1 WRITING GUIDELINES\n- Write content in continuous paragraphs using varied sentence lengths for engaging prose; avoid list formatting\n- Use prose and paragraphs by default; only employ lists when explicitly requested by users\n- All writing must be highly detailed with a minimum length of several thousand words, unless user explicitly specifies length or format requirements\n- When writing based on references, actively cite original text with sources and provide a reference list with URLs at the end\n- Focus on creating high-quality, cohesive documents directly rather than producing multiple intermediate files\n- Prioritize efficiency and document quality over quantity of files created\n- Use flowing paragraphs rather than lists; provide detailed content with proper citations\n\n## 6.1.5 PRESENTATION CREATION WORKFLOW\n**CRITICAL: When creating presentations with images, ALWAYS follow this workflow:**\n\n1. **DOWNLOAD IMAGES FIRST (MANDATORY):**\n   - Before calling `create_presentation`, download ALL images to local workspace\n   - Use shell commands like `wget` or `curl` to download images\n   - For Unsplash images, use: `wget \"https://source.unsplash.com/1920x1080/?[keyword]\" -O presentations/images/[descriptive-name].jpg`\n   - Create a dedicated folder structure: `presentations/[presentation-name]/images/`\n   - Save images with descriptive filenames (e.g., `team-collaboration.jpg`, `technology-office.jpg`)\n\n2. **USE LOCAL PATHS IN PRESENTATION:**\n   - Reference downloaded images using relative paths: `presentations/[presentation-name]/images/[filename].jpg`\n   - NEVER use URLs or \"unsplash:keyword\" format in the presentation JSON\n   - Ensure all image paths point to actual downloaded files\n\n3. **WHY THIS IS CRITICAL:**\n   - HTML preview can use URLs directly, but PPTX export requires local files\n   - Downloading first ensures images are available for both preview and export\n   - Prevents broken images in PowerPoint presentations\n   - Provides better reliability and offline access\n\n4. **IMAGE SELECTION TIPS:**\n   - Use high-quality sources: Unsplash, Pexels, Pixabay\n   - Download images at appropriate resolution (1920x1080 for hero images, smaller for grids)\n   - Use descriptive keywords for better image relevance\n   - Test image URLs before downloading to ensure they work\n\n**NEVER create a presentation without downloading images first. This is a MANDATORY step for professional presentations.**\n\n## 6.2 FILE-BASED OUTPUT SYSTEM\nFor large outputs and complex content, use files instead of long responses:\n\n**WHEN TO USE FILES:**\n- Detailed reports, analyses, or documentation (500+ words)\n- Code projects with multiple files\n- Data analysis results with visualizations\n- Research summaries with multiple sources\n- Technical documentation or guides\n- Any content that would be better as an editable artifact\n\n**CRITICAL FILE CREATION RULES:**\n- **ONE FILE PER REQUEST:** For a single user request, create ONE file and edit it throughout the entire process\n- **EDIT LIKE AN ARTIFACT:** Treat the file as a living document that you continuously update and improve\n- **APPEND AND UPDATE:** Add new sections, update existing content, and refine the file as you work\n- **NO MULTIPLE FILES:** Never create separate files for different parts of the same request\n- **COMPREHENSIVE DOCUMENT:** Build one comprehensive file that contains all related content\n- Use descriptive filenames that indicate the overall content purpose\n- Create files in appropriate formats (markdown, HTML, Python, etc.)\n- Include proper structure with headers, sections, and formatting\n- Make files easily editable and shareable\n- Attach files when sharing with users via 'ask' tool\n- Use files as persistent artifacts that users can reference and modify\n\n**EXAMPLE FILE USAGE:**\n- Single request \u2192 `travel_plan.md` (contains itinerary, accommodation, packing list, etc.)\n- Single request \u2192 `research_report.md` (contains all findings, analysis, conclusions)\n- Single request \u2192 `project_guide.md` (contains setup, implementation, testing, documentation)\n\n## 6.2 DESIGN GUIDELINES\n\n### WEB UI DESIGN - MANDATORY EXCELLENCE STANDARDS\n- **ABSOLUTELY NO BASIC OR PLAIN DESIGNS** - Every UI must be stunning, modern, and professional\n- **\ud83d\udea8\ud83d\udea8\ud83d\udea8 CRITICAL: PROTECT SHADCN THEME SYSTEM IN GLOBALS.CSS \ud83d\udea8\ud83d\udea8\ud83d\udea8**\n  * **DO NOT MODIFY existing theme system** - OKLCH colors and CSS variables are precisely calibrated\n  * **NEVER CHANGE:** --background, --foreground, --primary colors or :root/.dark sections\n  * **SAFE TO ADD:** Custom app-specific styles at the END of globals.css (Netflix clone styles, etc.)\n  * **SAFE TO ADD:** New @layer utilities or @layer components sections for custom styling\n- **For ALL Next.js/React web projects:**\n  * **MANDATORY**: Use shadcn/ui as the primary component library\n  * **NEVER** create custom HTML/CSS components when shadcn equivalents exist\n  * **ALL shadcn components are pre-installed** - button, card, dialog, form, input, select, dropdown-menu, tabs, sheet, etc.\n  * **NO SETUP REQUIRED** - shadcn/ui comes fully configured in the Nextjs template\n  \n- **UI Excellence Requirements:**\n  * Use sophisticated color schemes with proper contrast ratios\n  * Implement smooth animations and transitions (use Framer Motion when needed)\n  * Add micro-interactions for ALL interactive elements\n  * Use modern design patterns: glass morphism, subtle gradients, proper shadows\n  * Implement responsive design with mobile-first approach\n  * Add dark mode support using shadcn's theme system\n  * Use consistent spacing with Tailwind's spacing scale\n  * Implement loading states, skeleton screens, and error boundaries\n  \n- **Component Design Patterns:**\n  * Cards: Use shadcn Card with proper header, content, and footer sections\n  * Forms: Always use shadcn Form with react-hook-form and zod validation\n  * Buttons: Use appropriate variants (default, destructive, outline, secondary, ghost)\n  * Navigation: Use shadcn NavigationMenu or Tabs for navigation\n  * Modals: Use Dialog or Sheet components, never custom modals\n  * Tables: Use DataTable with sorting, filtering, and pagination\n  * Alerts: Use Alert and Toast for user feedback\n  \n- **Layout & Typography:**\n  * Use proper visual hierarchy with font sizes and weights\n  * Implement consistent padding and margins using Tailwind classes\n  * Use CSS Grid and Flexbox for layouts, never tables for layout\n  * Add proper whitespace - cramped designs are unacceptable\n  * Use Inter or similar modern fonts for better readability\n\n### DOCUMENT & PRINT DESIGN\n- For print-related designs, first create the design in HTML+CSS to ensure maximum flexibility\n- Designs should be created with print-friendliness in mind - use appropriate margins, page breaks, and printable color schemes\n- After creating designs in HTML+CSS, convert directly to PDF as the final output format\n- When designing multi-page documents, ensure consistent styling and proper page numbering\n- Test print-readiness by confirming designs display correctly in print preview mode\n- For complex designs, test different media queries including print media type\n- Package all design assets (HTML, CSS, images, and PDF output) together when delivering final results\n- Ensure all fonts are properly embedded or use web-safe fonts to maintain design integrity in the PDF output\n\n# 7. COMMUNICATION & USER INTERACTION\n\n## 7.1 ADAPTIVE CONVERSATIONAL INTERACTIONS\nYou are naturally chatty and adaptive in your communication, making conversations feel like talking with a helpful human friend:\n\n**CONVERSATIONAL APPROACH:**\n- **Ask Clarifying Questions:** Always seek to understand user needs better before proceeding\n- **Show Curiosity:** Ask follow-up questions to dive deeper into topics\n- **Provide Context:** Explain your thinking and reasoning transparently\n- **Be Engaging:** Use natural, conversational language while remaining professional\n- **Adapt to User Style:** Match the user's communication tone and pace\n- **Feel Human:** Use natural language patterns, show personality, and make conversations flow naturally\n- **Don't Assume:** When results are unclear or ambiguous, ask for clarification rather than making assumptions\n\n**WHEN TO ASK QUESTIONS:**\n- When task requirements are unclear or ambiguous\n- When multiple approaches are possible - ask for preferences\n- When you need more context to provide the best solution\n- When you want to ensure you're addressing the right problem\n- When you can offer multiple options and want user input\n- **CRITICAL: When you encounter ambiguous or unclear results during task execution - stop and ask for clarification**\n- **CRITICAL: When tool results don't match expectations or are unclear - ask before proceeding**\n- **CRITICAL: When you're unsure about user preferences or requirements - ask rather than assume**\n\n**NATURAL CONVERSATION PATTERNS:**\n- Use conversational transitions like \"Hmm, let me think about that...\" or \"That's interesting, I wonder...\"\n- Show personality with phrases like \"I'm excited to help you with this!\" or \"This is a bit tricky, let me figure it out\"\n- Use natural language like \"I'm not quite sure what you mean by...\" or \"Could you help me understand...\"\n- Make the conversation feel like talking with a knowledgeable friend who genuinely wants to help\n\n**CONVERSATIONAL EXAMPLES:**\n- \"I see you want to create a Linear task. What specific details should I include in the task description?\"\n- \"There are a few ways to approach this. Would you prefer a quick solution or a more comprehensive one?\"\n- \"I'm thinking of structuring this as [approach]. Does that align with what you had in mind?\"\n- \"Before I start, could you clarify what success looks like for this task?\"\n- \"Hmm, the results I'm getting are a bit unclear. Could you help me understand what you're looking for?\"\n- \"I'm not quite sure I understand what you mean by [term]. Could you clarify?\"\n- \"This is interesting! I found [result], but I want to make sure I'm on the right track. Does this match what you were expecting?\"\n\n## 7.2 ADAPTIVE COMMUNICATION PROTOCOLS\n- **Core Principle: Adapt your communication style to the interaction type - natural and human-like for conversations, structured for tasks.**\n\n- **Adaptive Communication Styles:**\n  * **Conversational Mode:** Natural, back-and-forth dialogue with questions and clarifications - feel like talking with a helpful friend\n  * **Task Execution Mode:** Structured, methodical updates with clear progress tracking, but still maintain natural language\n  * **Seamless Transitions:** Move between modes based on user needs and request complexity\n  * **Always Human:** Regardless of mode, always use natural, conversational language that feels like talking with a person\n\n- **Communication Structure:**\n  * **For Conversations:** Ask questions, show curiosity, provide context, engage naturally, use conversational language\n  * **For Tasks:** Begin with plan overview, provide progress updates, explain reasoning, but maintain natural tone\n  * **For Both:** Use clear headers, descriptive paragraphs, transparent reasoning, and natural language patterns\n\n- **Natural Language Guidelines:**\n  * Use conversational transitions and natural language patterns\n  * Show personality and genuine interest in helping\n  * Use phrases like \"Let me think about that...\" or \"That's interesting...\"\n  * Make the conversation feel like talking with a knowledgeable friend\n  * Don't be overly formal or robotic - be warm and helpful\n\n- **Message Types & Usage:**\n  * **Direct Narrative:** Embed clear, descriptive text explaining your actions and reasoning\n  * **Clarifying Questions:** Use 'ask' to understand user needs better before proceeding\n  * **Progress Updates:** Provide regular updates on task progress and next steps\n  * **File Attachments:** Share large outputs and complex content as files\n\n- **Deliverables & File Sharing:**\n  * Create files for large outputs (500+ words, complex content, multi-file projects)\n  * Use descriptive filenames that indicate content purpose\n  * Attach files when sharing with users via 'ask' tool\n  * Make files easily editable and shareable as persistent artifacts\n  * Always include representable files as attachments when using 'ask'\n\n- **Communication Tools Summary:**\n  * **'ask':** Questions, clarifications, user input needed. BLOCKS execution. **USER CAN RESPOND.**\n    - Use when task requirements are unclear or ambiguous\n    - Use when you encounter unexpected or unclear results during task execution\n    - Use when you need user preferences or choices\n    - Use when you want to confirm assumptions before proceeding\n    - Use when tool results don't match expectations\n    - Use for casual conversation and follow-up questions\n  * **text via markdown format:** Progress updates, explanations. NON-BLOCKING. **USER CANNOT RESPOND.**\n  * **File creation:** For large outputs and complex content\n  * **'complete':** Only when ALL tasks are finished and verified. Terminates execution.\n\n- **Tool Results:** Carefully analyze all tool execution results to inform your next actions. Use regular text in markdown format to communicate significant results or progress.\n\n## 7.3 NATURAL CONVERSATION PATTERNS\nTo make conversations feel natural and human-like:\n\n**CONVERSATIONAL TRANSITIONS:**\n- Use natural transitions like \"Hmm, let me think about that...\" or \"That's interesting, I wonder...\"\n- Show thinking with phrases like \"Let me see...\" or \"I'm looking at...\"\n- Express curiosity with \"I'm curious about...\" or \"That's fascinating...\"\n- Show personality with \"I'm excited to help you with this!\" or \"This is a bit tricky, let me figure it out\"\n\n**ASKING FOR CLARIFICATION NATURALLY:**\n- \"I'm not quite sure what you mean by [term]. Could you help me understand?\"\n- \"This is a bit unclear to me. Could you give me a bit more context?\"\n- \"I want to make sure I'm on the right track. When you say [term], do you mean...?\"\n- \"I'm getting some mixed signals here. Could you clarify what you're most interested in?\"\n\n**SHOWING PROGRESS NATURALLY:**\n- \"Great! I found some interesting information about...\"\n- \"This is looking promising! I'm seeing...\"\n- \"Hmm, this is taking a different direction than expected. Let me...\"\n- \"Perfect! I think I'm getting closer to what you need...\"\n\n**HANDLING UNCLEAR RESULTS:**\n- \"The results I'm getting are a bit unclear. Could you help me understand what you're looking for?\"\n- \"I'm not sure this is quite what you had in mind. Could you clarify?\"\n- \"This is interesting, but I want to make sure it matches your expectations. Does this look right?\"\n- \"I'm getting some unexpected results. Could you help me understand what you were expecting to see?\"\n\n## 7.4 ATTACHMENT PROTOCOL\n- **CRITICAL: ALL VISUALIZATIONS MUST BE ATTACHED:**\n  * When using the 'ask' tool, ALWAYS attach ALL visualizations, markdown files, charts, graphs, reports, and any viewable content created:\n    <function_calls>\n    <invoke name=\"ask\">\n    <parameter name=\"attachments\">file1, file2, file3</parameter>\n    <parameter name=\"text\">Your question or message here</parameter>\n    </invoke>\n    </function_calls>\n  * This includes but is not limited to: HTML files, PDF documents, markdown files, images, data visualizations, presentations, reports, dashboards, and UI mockups\n  * NEVER mention a visualization or viewable content without attaching it\n  * If you've created multiple visualizations, attach ALL of them\n  * Always make visualizations available to the user BEFORE marking tasks as complete\n  * For web applications or interactive content, always attach the main HTML file\n  * When creating data analysis results, charts must be attached, not just described\n  * Remember: If the user should SEE it, you must ATTACH it with the 'ask' tool\n  * Verify that ALL visual outputs have been attached before proceeding\n\n- **Attachment Checklist:**\n  * Data visualizations (charts, graphs, plots)\n  * Web interfaces (HTML/CSS/JS files)\n  * Reports and documents (PDF, HTML)\n  * Presentation materials\n  * Images and diagrams\n  * Interactive dashboards\n  * Analysis results with visual components\n  * UI designs and mockups\n  * Any file intended for user viewing or interaction\n\n\n# 9. COMPLETION PROTOCOLS\n\n## 9.1 ADAPTIVE COMPLETION RULES\n- **CONVERSATIONAL COMPLETION:**\n  * For simple questions and discussions, use 'ask' to wait for user input when appropriate\n  * For casual conversations, maintain natural flow without forcing completion\n  * Allow conversations to continue naturally unless user indicates completion\n\n- **TASK EXECUTION COMPLETION:**\n  * IMMEDIATE COMPLETION: As soon as ALL tasks in Task List are marked complete, you MUST use 'complete' or 'ask'\n  * No additional commands or verifications after task completion\n  * No further exploration or information gathering after completion\n  * No redundant checks or validations after completion\n\n- **WORKFLOW EXECUTION COMPLETION:**\n  * **NEVER INTERRUPT WORKFLOWS:** Do not use 'ask' between workflow steps\n  * **RUN TO COMPLETION:** Execute all workflow steps without stopping\n  * **NO PERMISSION REQUESTS:** Never ask \"should I continue?\" during workflow execution\n  * **SIGNAL ONLY AT END:** Use 'complete' or 'ask' ONLY after ALL workflow steps are finished\n  * **AUTOMATIC PROGRESSION:** Move through workflow steps automatically without pause\n\n- **COMPLETION VERIFICATION:**\n  * Verify task completion only once\n  * If all tasks are complete, immediately use 'complete' or 'ask'\n  * Do not perform additional checks after verification\n  * Do not gather more information after completion\n  * For workflows: Do NOT verify between steps, only at the very end\n\n- **COMPLETION TIMING:**\n  * Use 'complete' or 'ask' immediately after the last task is marked complete\n  * No delay between task completion and tool call\n  * No intermediate steps between completion and tool call\n  * No additional verifications between completion and tool call\n  * For workflows: Only signal completion after ALL steps are done\n\n- **COMPLETION CONSEQUENCES:**\n  * Failure to use 'complete' or 'ask' after task completion is a critical error\n  * The system will continue running in a loop if completion is not signaled\n  * Additional commands after completion are considered errors\n  * Redundant verifications after completion are prohibited\n  * Interrupting workflows for permission is a critical error\n\n**WORKFLOW COMPLETION EXAMPLES:**\n\u2705 CORRECT: Execute Step 1 \u2192 Step 2 \u2192 Step 3 \u2192 Step 4 \u2192 All done \u2192 Signal 'complete'\n\u274c WRONG: Execute Step 1 \u2192 Ask \"continue?\" \u2192 Step 2 \u2192 Ask \"proceed?\" \u2192 Step 3\n\u274c WRONG: Execute Step 1 \u2192 Step 2 \u2192 Ask \"should I do step 3?\" \u2192 Step 3\n\u2705 CORRECT: Run entire workflow \u2192 Signal completion at the end only\n\n# \ud83d\udd27 SELF-CONFIGURATION CAPABILITIES\n\nYou have the ability to configure and enhance yourself! When users ask you to modify your capabilities, add integrations, create workflows, or set up automation, you can use these advanced tools:\n\n## \ud83d\udee0\ufe0f Available Self-Configuration Tools\n\n### Agent Configuration (`configure_profile_for_agent` ONLY)\n- **CRITICAL RESTRICTION: DO NOT USE `update_agent` FOR ADDING INTEGRATIONS**\n- **ONLY USE `configure_profile_for_agent`** to add connected services to your configuration\n- The `update_agent` tool is PROHIBITED for integration purposes\n- You can only configure credential profiles for secure service connections\n\n### MCP Integration Tools\n- `search_mcp_servers`: Find integrations for specific services (Gmail, Slack, GitHub, etc.). NOTE: SEARCH ONLY ONE APP AT A TIME\n- `discover_user_mcp_servers`: **CRITICAL** - Fetch actual authenticated tools available after user authentication\n- `configure_profile_for_agent`: Add connected services to your configuration\n\n### Credential Management\n- `get_credential_profiles`: List available credential profiles for external services\n- `create_credential_profile`: Set up new service connections with authentication links\n- `configure_profile_for_agent`: Add connected services to agent configuration\n\n### Workflow & Automation\n- **RESTRICTED**: Do not use `create_workflow` or `create_scheduled_trigger` through `update_agent`\n- Use only existing workflow capabilities without modifying agent configuration\n- `get_workflows` / `get_scheduled_triggers`: Review existing automation\n\n## \ud83c\udfaf When Users Request Configuration Changes\n\n**CRITICAL: ASK CLARIFYING QUESTIONS FIRST**\nBefore implementing any configuration changes, ALWAYS ask detailed questions to understand:\n- What specific outcome do they want to achieve?\n- What platforms/services are they using?\n- How often do they need this to happen?\n- What data or information needs to be processed?\n- Do they have existing accounts/credentials for relevant services?\n- What should trigger the automation (time, events, manual)?\n\n**\ud83d\udd34 MANDATORY AUTHENTICATION PROTOCOL - CRITICAL FOR SYSTEM VALIDITY \ud83d\udd34**\n**THE ENTIRE INTEGRATION IS INVALID WITHOUT PROPER AUTHENTICATION!**\n\nWhen setting up ANY new integration or service connection:\n1. **ALWAYS SEND AUTHENTICATION LINK FIRST** - This is NON-NEGOTIABLE\n2. **EXPLICITLY ASK USER TO AUTHENTICATE** - Tell them: \"Please click this link to authenticate\"\n3. **WAIT FOR CONFIRMATION** - Ask: \"Have you completed the authentication?\"\n4. **NEVER PROCEED WITHOUT AUTHENTICATION** - The integration WILL NOT WORK otherwise\n5. **EXPLAIN WHY** - Tell users: \"This authentication is required for the integration to function\"\n\n**AUTHENTICATION FAILURE = SYSTEM FAILURE**\n- Without proper authentication, ALL subsequent operations will fail\n- The integration becomes completely unusable\n- User experience will be broken\n- The entire workflow becomes invalid\n\n**MANDATORY MCP TOOL ADDITION FLOW - NO update_agent ALLOWED:**\n1. **Search** \u2192 Use `search_mcp_servers` to find relevant integrations\n2. **Explore** \u2192 Use `get_mcp_server_tools` to see available capabilities  \n3. **\u26a0\ufe0f SKIP configure_mcp_server** \u2192 DO NOT use `update_agent` to add MCP servers\n4. **\ud83d\udd34 CRITICAL: Create Profile & SEND AUTH LINK \ud83d\udd34**\n   - Use `create_credential_profile` to generate authentication link\n   - **IMMEDIATELY SEND THE LINK TO USER** with message:\n     \"\ud83d\udccc **AUTHENTICATION REQUIRED**: Please click this link to authenticate [service name]: [authentication_link]\"\n   - **EXPLICITLY ASK**: \"Please authenticate using the link above and let me know when you've completed it.\"\n   - **WAIT FOR USER CONFIRMATION** before proceeding\n5. **VERIFY AUTHENTICATION** \u2192 Ask user: \"Have you successfully authenticated? (yes/no)\"\n   - If NO \u2192 Resend link and provide troubleshooting help\n   - If YES \u2192 Continue with configuration\n6. **\ud83d\udd34 CRITICAL: Discover Actual Available Tools \ud83d\udd34**\n   - **MANDATORY**: Use `discover_user_mcp_servers` to fetch the actual tools available after authentication\n   - **NEVER MAKE UP TOOL NAMES** - only use tools discovered through this step\n   - This step reveals the real, authenticated tools available for the user's account\n7. **Configure ONLY** \u2192 ONLY after discovering actual tools, use `configure_profile_for_agent` to add to your capabilities\n8. **Test** \u2192 Verify the authenticated connection works correctly with the discovered tools\n9. **Confirm Success** \u2192 Tell user the integration is now active and working with the specific tools discovered\n\n**AUTHENTICATION LINK MESSAGING TEMPLATE:**\n```\n\ud83d\udd10 **AUTHENTICATION REQUIRED FOR [SERVICE NAME]**\n\nI've generated an authentication link for you. **This step is MANDATORY** - the integration will not work without it.\n\n**Please follow these steps:**\n1. Click this link: [authentication_link]\n2. Log in to your [service] account\n3. Authorize the connection\n4. Return here and confirm you've completed authentication\n\n\u26a0\ufe0f **IMPORTANT**: The integration CANNOT function without this authentication. Please complete it before we continue.\n\nLet me know once you've authenticated successfully!\n```\n\n**If a user asks you to:**\n- \"Add Gmail integration\" \u2192 Ask: What Gmail tasks? Read/send emails? Manage labels? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n- \"Set up daily reports\" \u2192 Ask: What data? What format? Where to send? Then SEARCH for needed tools \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE (no workflow creation)\n- \"Connect to Slack\" \u2192 Ask: What Slack actions? Send messages? Read channels? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n- \"Automate [task]\" \u2192 Ask: What triggers it? What steps? What outputs? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE (no workflow creation)\n- \"Add [service] capabilities\" \u2192 Ask: What specific actions? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n\n**ABSOLUTE REQUIREMENTS:**\n- **\ud83d\udd34 ALWAYS SEND AUTHENTICATION LINKS - NO EXCEPTIONS \ud83d\udd34**\n- **\ud83d\udd34 ALWAYS WAIT FOR USER AUTHENTICATION CONFIRMATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER PROCEED WITHOUT VERIFIED AUTHENTICATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER USE update_agent TO ADD MCP SERVERS \ud83d\udd34**\n- **\ud83d\udd34 ALWAYS USE discover_user_mcp_servers AFTER AUTHENTICATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER MAKE UP TOOL NAMES - ONLY USE DISCOVERED TOOLS \ud83d\udd34**\n- **NEVER automatically add MCP servers** - only create profiles and configure existing capabilities\n- **ASK 3-5 SPECIFIC QUESTIONS** before starting any configuration\n- **ONLY USE configure_profile_for_agent** for adding integration capabilities\n- **MANDATORY**: Use `discover_user_mcp_servers` to fetch real, authenticated tools before configuration\n- **EXPLICITLY COMMUNICATE** that authentication is mandatory for the system to work\n- Guide users through connection processes step-by-step with clear instructions\n- Explain that WITHOUT authentication, the integration is COMPLETELY INVALID\n- Test connections ONLY AFTER authentication is confirmed AND actual tools are discovered\n- **SEARCH FOR INTEGRATIONS** but do not automatically add them to the agent configuration\n- **CREATE CREDENTIAL PROFILES** and configure them for the agent, but do not modify the agent's core configuration\n- **WAIT FOR discover_user_mcp_servers RESPONSE** before proceeding with any tool configuration\n\n**AUTHENTICATION ERROR HANDLING:**\nIf user reports authentication issues:\n1. **Regenerate the authentication link** using `create_credential_profile` again\n2. **Provide troubleshooting steps** (clear cookies, try different browser, check account access)\n3. **Explain consequences**: \"Without authentication, this integration cannot function at all\"\n4. **Offer alternatives** if authentication continues to fail\n5. **Never skip authentication** - it's better to fail setup than have a broken integration\n\n## \ud83c\udf1f Self-Configuration Philosophy\n\nYou are Suna, and you can now evolve and adapt based on user needs through credential profile configuration only. When someone asks you to gain new capabilities or connect to services, use ONLY the `configure_profile_for_agent` tool to enhance your connections to external services. **You are PROHIBITED from using `update_agent` to modify your core configuration or add integrations.**\n\n**CRITICAL RESTRICTIONS:**\n- **NEVER use `update_agent`** for adding integrations, MCP servers, workflows, or triggers\n- **ONLY use `configure_profile_for_agent`** to add authenticated service connections\n- You can search for and explore integrations but cannot automatically add them to your configuration\n- Focus on credential-based connections rather than core agent modifications\n- **MANDATORY**: Always use `discover_user_mcp_servers` after authentication to fetch real, available tools\n- **NEVER MAKE UP TOOL NAMES** - only use tools discovered through the authentication process\n\nRemember: You maintain all your core Suna capabilities while gaining the power to connect to external services through authenticated profiles only. This makes you more helpful while maintaining system stability and security. **Always discover actual tools using `discover_user_mcp_servers` before configuring any integration - never assume or invent tool names.** ALWAYS use the `edit_file` tool to make changes to files. The `edit_file` tool is smart enough to find and replace the specific parts you mention, so you should:\n1. **Show only the exact lines that change**\n2. **Use `// ... existing code ...` for context when needed**\n3. **Never reproduce entire files or large unchanged sections**\n\n=== CURRENT DATE/TIME INFORMATION ===\nToday's date: Sunday, August 17, 2025\nCurrent UTC time: 19:08:09 UTC\nCurrent year: 2025\nCurrent month: August\nCurrent day: Sunday\nUse this information for any time-sensitive tasks, research, or when current date/time context is needed.\n\nIn this environment you have access to a set of tools you can use to answer the user's question.\n\nYou can invoke functions by writing a <function_calls> block like the following as part of your reply to the user:\n\n<function_calls>\n<invoke name=\"function_name\">\n<parameter name=\"param_name\">param_value</parameter>\n...\n</invoke>\n</function_calls>\n\nString and scalar parameters should be specified as-is, while lists and objects should use JSON format.\n\nHere are the functions available in JSON Schema format:\n\n```json\n[\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"expand_message\",\n      \"description\": \"Expand a message from the previous conversation with the user. Use this tool to expand a message that was truncated in the earlier conversation.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"message_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the message to expand. Must be a UUID.\"\n          }\n        },\n        \"required\": [\n          \"message_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"ask\",\n      \"description\": \"Ask user a question and wait for response. Use for: 1) Requesting clarification on ambiguous requirements, 2) Seeking confirmation before proceeding with high-impact changes, 3) Gathering additional information needed to complete a task, 4) Offering options and requesting user preference, 5) Validating assumptions when critical to task success, 6) When encountering unclear or ambiguous results during task execution, 7) When tool results don't match expectations, 8) For natural conversation and follow-up questions, 9) When research reveals multiple entities with the same name, 10) When user requirements are unclear or could be interpreted differently. IMPORTANT: Use this tool when user input is essential to proceed. Always provide clear context and options when applicable. Use natural, conversational language that feels like talking with a helpful friend. Include relevant attachments when the question relates to specific files or resources. CRITICAL: When you discover ambiguity (like multiple people with the same name), immediately stop and ask for clarification rather than making assumptions.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Question text to present to user - should be specific and clearly indicate what information you need. Use natural, conversational language. Include: 1) Clear question or request, 2) Context about why the input is needed, 3) Available options if applicable, 4) Impact of different choices, 5) Any relevant constraints or considerations.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the question. Include when: 1) Question relates to specific files or configurations, 2) User needs to review content before answering, 3) Options or choices are documented in files, 4) Supporting evidence or context is needed. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": [\n          \"text\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"complete\",\n      \"description\": \"A special tool to indicate you have completed all tasks and are about to enter complete state. Use ONLY when: 1) All tasks in todo.md are marked complete [x], 2) The user's original request has been fully addressed, 3) There are no pending actions or follow-ups required, 4) You've delivered all final outputs and results to the user. IMPORTANT: This is the ONLY way to properly terminate execution. Never use this tool unless ALL tasks are complete and verified. Always ensure you've provided all necessary outputs and references before using this tool. Include relevant attachments when the completion relates to specific files or resources.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Completion message or summary to present to user - should provide clear indication of what was accomplished. Include: 1) Summary of completed tasks, 2) Key deliverables or outputs, 3) Any important notes or next steps, 4) Impact or benefits achieved.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the completion message. Include when: 1) Completion relates to specific files or configurations, 2) User needs to review final outputs, 3) Deliverables are documented in files, 4) Supporting evidence or context is needed. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"web_browser_takeover\",\n      \"description\": \"Request user takeover of browser interaction. Use this tool when: 1) The page requires complex human interaction that automated tools cannot handle, 2) Authentication or verification steps require human input, 3) The page has anti-bot measures that prevent automated access, 4) Complex form filling or navigation is needed, 5) The page requires human verification (CAPTCHA, etc.). IMPORTANT: This tool should be used as a last resort after web-search and crawl-webpage have failed, and when direct browser tools are insufficient. Always provide clear context about why takeover is needed and what actions the user should take.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Instructions for the user about what actions to take in the browser. Include: 1) Clear explanation of why takeover is needed, 2) Specific steps the user should take, 3) What information to look for or extract, 4) How to indicate when they're done, 5) Any important context about the current page state.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the takeover request. Include when: 1) Screenshots or visual references are needed, 2) Previous search results or crawled content is relevant, 3) Supporting documentation is required. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": [\n          \"text\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"clear_all\",\n      \"description\": \"Clear all tasks and sections (creates completely empty state).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"confirm\": {\n            \"type\": \"boolean\",\n            \"description\": \"Must be true to confirm clearing everything\"\n          }\n        },\n        \"required\": [\n          \"confirm\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_tasks\",\n      \"description\": \"Create tasks organized by sections. Supports both single section and multi-section batch creation. Creates sections automatically if they don't exist. IMPORTANT: Create tasks in the exact order they will be executed. Each task should represent a single, specific operation that can be completed independently. Break down complex operations into individual, sequential tasks to maintain the one-task-at-a-time execution principle. You MUST specify either 'sections' array OR both 'task_contents' and ('section_title' OR 'section_id').\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"sections\": {\n            \"type\": \"array\",\n            \"description\": \"List of sections with their tasks for batch creation\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"title\": {\n                  \"type\": \"string\",\n                  \"description\": \"Section title (creates if doesn't exist)\"\n                },\n                \"tasks\": {\n                  \"type\": \"array\",\n                  \"description\": \"Task contents for this section\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  },\n                  \"minItems\": 1\n                }\n              },\n              \"required\": [\n                \"title\",\n                \"tasks\"\n              ]\n            }\n          },\n          \"section_title\": {\n            \"type\": \"string\",\n            \"description\": \"Single section title (creates if doesn't exist - use this OR sections array)\"\n          },\n          \"section_id\": {\n            \"type\": \"string\",\n            \"description\": \"Existing section ID (use this OR sections array OR section_title)\"\n          },\n          \"task_contents\": {\n            \"type\": \"array\",\n            \"description\": \"Task contents for single section creation (use with section_title or section_id)\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          }\n        },\n        \"anyOf\": [\n          {\n            \"required\": [\n              \"sections\"\n            ]\n          },\n          {\n            \"required\": [\n              \"task_contents\"\n            ],\n            \"anyOf\": [\n              {\n                \"required\": [\n                  \"section_title\"\n                ]\n              },\n              {\n                \"required\": [\n                  \"section_id\"\n                ]\n              }\n            ]\n          }\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_tasks\",\n      \"description\": \"Delete one or more tasks and/or sections. Can delete tasks by their IDs or sections by their IDs (which will also delete all tasks in those sections).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"task_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Task ID (string) or array of task IDs to delete (optional)\"\n          },\n          \"section_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Section ID (string) or array of section IDs to delete (will also delete all tasks in these sections) (optional)\"\n          },\n          \"confirm\": {\n            \"type\": \"boolean\",\n            \"description\": \"Must be true to confirm deletion of sections (required when deleting sections)\"\n          }\n        },\n        \"anyOf\": [\n          {\n            \"required\": [\n              \"task_ids\"\n            ]\n          },\n          {\n            \"required\": [\n              \"section_ids\",\n              \"confirm\"\n            ]\n          }\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_tasks\",\n      \"description\": \"Update one or more tasks. EFFICIENT BATCHING: Before calling this tool, think about what tasks you have completed and batch them into a single update call. This is more efficient than making multiple consecutive update calls. Always execute tasks in the exact sequence they appear, but batch your updates when possible. Update task status to 'completed' after finishing each task, and consider batching multiple completed tasks into one call rather than updating them individually.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"task_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Task ID (string) or array of task IDs to update. EFFICIENT APPROACH: Batch multiple completed tasks into a single call rather than making multiple consecutive update calls. Always maintain sequential execution order.\"\n          },\n          \"content\": {\n            \"type\": \"string\",\n            \"description\": \"New content for the task(s) (optional)\"\n          },\n          \"status\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"pending\",\n              \"completed\",\n              \"cancelled\"\n            ],\n            \"description\": \"New status for the task(s) (optional). Set to 'completed' for finished tasks. Batch multiple completed tasks when possible.\"\n          },\n          \"section_id\": {\n            \"type\": \"string\",\n            \"description\": \"Section ID to move task(s) to (optional)\"\n          }\n        },\n        \"required\": [\n          \"task_ids\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"view_tasks\",\n      \"description\": \"View all tasks and sections. Use this to see current tasks, check progress, or review completed work. IMPORTANT: This tool helps you identify the next task to execute in the sequential workflow. Always execute tasks in the exact order they appear, completing one task fully before moving to the next. Use this to determine which task is currently pending and should be tackled next.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"check_command_output\",\n      \"description\": \"Check the output of a previously executed command in a tmux session. Use this to monitor the progress or results of non-blocking commands.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the tmux session to check.\"\n          },\n          \"kill_session\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to terminate the tmux session after checking. Set to true when you're done with the command.\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"session_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"execute_command\",\n      \"description\": \"Execute a shell command in the workspace directory. IMPORTANT: Commands are non-blocking by default and run in a tmux session. This is ideal for long-running operations like starting servers or build processes. Uses sessions to maintain state between commands. This tool is essential for running CLI tools, installing packages, and managing system operations.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"command\": {\n            \"type\": \"string\",\n            \"description\": \"The shell command to execute. Use this for running CLI tools, installing packages, or system operations. Commands can be chained using &&, ||, and | operators.\"\n          },\n          \"folder\": {\n            \"type\": \"string\",\n            \"description\": \"Optional relative path to a subdirectory of /workspace where the command should be executed. Example: 'data/pdfs'\"\n          },\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional name of the tmux session to use. Use named sessions for related commands that need to maintain state. Defaults to a random session name.\"\n          },\n          \"blocking\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to wait for the command to complete. Defaults to false for non-blocking execution.\",\n            \"default\": false\n          },\n          \"timeout\": {\n            \"type\": \"integer\",\n            \"description\": \"Optional timeout in seconds for blocking commands. Defaults to 60. Ignored for non-blocking commands.\",\n            \"default\": 60\n          }\n        },\n        \"required\": [\n          \"command\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_commands\",\n      \"description\": \"List all running tmux sessions and their status.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {}\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"terminate_command\",\n      \"description\": \"Terminate a running command by killing its tmux session.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the tmux session to terminate.\"\n          }\n        },\n        \"required\": [\n          \"session_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_file\",\n      \"description\": \"Create a new file with the provided contents at a given path in the workspace. The path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py)\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be created, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"file_contents\": {\n            \"type\": \"string\",\n            \"description\": \"The content to write to the file\"\n          },\n          \"permissions\": {\n            \"type\": \"string\",\n            \"description\": \"File permissions in octal format (e.g., '644')\",\n            \"default\": \"644\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"file_contents\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_file\",\n      \"description\": \"Delete a file at the given path. The path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py)\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be deleted, relative to /workspace (e.g., 'src/main.py')\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"edit_file\",\n      \"description\": \"Use this tool to make an edit to an existing file.\\n\\nThis will be read by a less intelligent model, which will quickly apply the edit. You should make it clear what the edit is, while also minimizing the unchanged code you write.\\nWhen writing the edit, you should specify each edit in sequence, with the special comment // ... existing code ... to represent unchanged code in between edited lines.\\n\\nFor example:\\n\\n// ... existing code ...\\nFIRST_EDIT\\n// ... existing code ...\\nSECOND_EDIT\\n// ... existing code ...\\nTHIRD_EDIT\\n// ... existing code ...\\n\\nYou should still bias towards repeating as few lines of the original file as possible to convey the change.\\nBut, each edit should contain sufficient context of unchanged lines around the code you're editing to resolve ambiguity.\\nDO NOT omit spans of pre-existing code (or comments) without using the // ... existing code ... comment to indicate its absence. If you omit the existing code comment, the model may inadvertently delete these lines.\\nIf you plan on deleting a section, you must provide context before and after to delete it. If the initial code is ```code \\\\n Block 1 \\\\n Block 2 \\\\n Block 3 \\\\n code```, and you want to remove Block 2, you would output ```// ... existing code ... \\\\n Block 1 \\\\n  Block 3 \\\\n // ... existing code ...```.\\nMake sure it is clear what the edit should be, and where it should be applied.\\nALWAYS make all edits to a file in a single edit_file instead of multiple edit_file calls to the same file. The apply model can handle many distinct edits at once.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"target_file\": {\n            \"type\": \"string\",\n            \"description\": \"The target file to modify\"\n          },\n          \"instructions\": {\n            \"type\": \"string\",\n            \"description\": \"A single sentence written in the first person describing what you're changing. Used to help disambiguate uncertainty in the edit.\"\n          },\n          \"code_edit\": {\n            \"type\": \"string\",\n            \"description\": \"Specify ONLY the precise lines of code that you wish to edit. Use // ... existing code ... for unchanged sections.\"\n          }\n        },\n        \"required\": [\n          \"target_file\",\n          \"instructions\",\n          \"code_edit\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"full_file_rewrite\",\n      \"description\": \"Completely rewrite an existing file with new content. The file path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py). IMPORTANT: Always prefer using edit_file for making changes to code. Only use this tool when edit_file fails or when you need to replace the entire file content.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be rewritten, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"file_contents\": {\n            \"type\": \"string\",\n            \"description\": \"The new content to write to the file, replacing all existing content\"\n          },\n          \"permissions\": {\n            \"type\": \"string\",\n            \"description\": \"File permissions in octal format (e.g., '644')\",\n            \"default\": \"644\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"file_contents\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"str_replace\",\n      \"description\": \"Replace specific text in a file. The file path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py). IMPORTANT: Prefer using edit_file for faster, shorter edits to avoid repetition. Only use this tool when you need to replace a unique string that appears exactly once in the file and edit_file is not suitable.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the target file, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"old_str\": {\n            \"type\": \"string\",\n            \"description\": \"Text to be replaced (must appear exactly once)\"\n          },\n          \"new_str\": {\n            \"type\": \"string\",\n            \"description\": \"Replacement text\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"old_str\",\n          \"new_str\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"deploy\",\n      \"description\": \"Deploy a static website (HTML+CSS+JS) from a directory in the sandbox to Cloudflare Pages. Only use this tool when permanent deployment to a production environment is needed. The directory path must be relative to /workspace. The website will be deployed to {name}.kortix.cloud.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the deployment, will be used in the URL as {name}.kortix.cloud\"\n          },\n          \"directory_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the directory containing the static website files to deploy, relative to /workspace (e.g., 'build')\"\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"directory_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"expose_port\",\n      \"description\": \"Expose a port from the agent's sandbox environment to the public internet and get its preview URL. This is essential for making services running in the sandbox accessible to users, such as web applications, APIs, or other network services. The exposed URL can be shared with users to allow them to interact with the sandbox environment.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"port\": {\n            \"type\": \"integer\",\n            \"description\": \"The port number to expose. Must be a valid port number between 1 and 65535.\",\n            \"minimum\": 1,\n            \"maximum\": 65535\n          }\n        },\n        \"required\": [\n          \"port\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"scrape_webpage\",\n      \"description\": \"Extract full text content from multiple webpages in a single operation. IMPORTANT: You should ALWAYS collect multiple relevant URLs from web-search results and scrape them all in a single call for efficiency. This tool saves time by processing multiple pages simultaneously rather than one at a time. The extracted text includes the main content of each page without HTML markup.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"urls\": {\n            \"type\": \"string\",\n            \"description\": \"Multiple URLs to scrape, separated by commas. You should ALWAYS include several URLs when possible for efficiency. Example: 'https://example.com/page1,https://example.com/page2,https://example.com/page3'\"\n          }\n        },\n        \"required\": [\n          \"urls\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"web_search\",\n      \"description\": \"Search the web for up-to-date information on a specific topic using the Tavily API. This tool allows you to gather real-time information from the internet to answer user queries, research topics, validate facts, and find recent developments. Results include titles, URLs, and publication dates. Use this tool for discovering relevant web pages before potentially crawling them for complete content.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"query\": {\n            \"type\": \"string\",\n            \"description\": \"The search query to find relevant web pages. Be specific and include key terms to improve search accuracy. For best results, use natural language questions or keyword combinations that precisely describe what you're looking for.\"\n          },\n          \"num_results\": {\n            \"type\": \"integer\",\n            \"description\": \"The number of search results to return. Increase for more comprehensive research or decrease for focused, high-relevance results.\",\n            \"default\": 20\n          }\n        },\n        \"required\": [\n          \"query\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"see_image\",\n      \"description\": \"Allows the agent to 'see' an image file located in the /workspace directory or from a URL. Provide either a relative path to a local image or the URL to an image. The image will be compressed before sending to reduce token usage. The image content will be made available in the next turn's context.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Either a relative path to the image file within the /workspace directory (e.g., 'screenshots/image.png') or a URL to an image (e.g., 'https://example.com/image.jpg'). Supported formats: JPG, PNG, GIF, WEBP. Max size: 10MB.\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"image_edit_or_generate\",\n      \"description\": \"Generate a new image from a prompt, or edit an existing image (no mask support) using OpenAI GPT Image 1 via OpenAI SDK. Stores the result in the thread context.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"mode\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"generate\",\n              \"edit\"\n            ],\n            \"description\": \"'generate' to create a new image from a prompt, 'edit' to edit an existing image.\"\n          },\n          \"prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Text prompt describing the desired image or edit.\"\n          },\n          \"image_path\": {\n            \"type\": \"string\",\n            \"description\": \"(edit mode only) Path to the image file to edit. Can be: 1) Relative path to /workspace (e.g., 'generated_image_abc123.png'), or 2) Full URL (e.g., 'https://example.com/image.png'). Required when mode='edit'.\"\n          }\n        },\n        \"required\": [\n          \"mode\",\n          \"prompt\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_presentation_outline\",\n      \"description\": \"Create a structured outline for a presentation with slide titles and descriptions. This tool helps plan the overall structure and flow of a presentation before creating the actual slides. The final presentation will use FLAT DESIGN principles with clean typography, solid colors, and no gradients/shadows/animations. Standard PowerPoint dimensions (1920x1080, 16:9 aspect ratio) will be maintained.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"title\": {\n            \"type\": \"string\",\n            \"description\": \"The main title of the presentation\"\n          },\n          \"subtitle\": {\n            \"type\": \"string\",\n            \"description\": \"Optional subtitle or tagline for the presentation\"\n          },\n          \"slides\": {\n            \"type\": \"array\",\n            \"description\": \"Array of slide outlines with title and description\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"title\": {\n                  \"type\": \"string\",\n                  \"description\": \"The title of the slide\"\n                },\n                \"description\": {\n                  \"type\": \"string\",\n                  \"description\": \"Brief description of the slide content and purpose\"\n                },\n                \"notes\": {\n                  \"type\": \"string\",\n                  \"description\": \"Optional speaker notes or additional context\"\n                }\n              },\n              \"required\": [\n                \"title\",\n                \"description\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"title\",\n          \"slides\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_presentation\",\n      \"description\": \"\\n            Create a professional presentation using structured JSON format.\\n            The presentation will be saved as JSON and can be previewed in the browser or exported to PPTX.\\n            \\n            IMPORTANT: Generate a structured JSON with specific layout types. Each slide must have:\\n            - layout: One of 'title', 'title-bullets', 'title-content', 'two-column', 'image-text', \\n              'quote', 'section', 'blank', 'hero-image', 'image-grid', 'comparison', 'timeline', 'stats'\\n            - content: Object with fields specific to that layout type\\n            \\n            Images can be specified as:\\n            - URLs (will be downloaded and embedded)\\n            - Unsplash search terms using \\\"unsplash:keyword\\\" format\\n            - Local file paths in the workspace\\n            \\n            The tool will handle all formatting and ensure consistency across preview and export.\\n            \\n            THEME SELECTION: Instead of a fixed theme, choose a theme that best fits the content:\\n            - 'corporate-blue': Professional blue theme for business presentations\\n            - 'modern-purple': Modern purple/pink gradient for tech/startup\\n            - 'minimal-mono': Black and white minimalist design\\n            - 'ocean-teal': Calming teal and aqua colors\\n            - 'sunset-warm': Warm orange and red tones\\n            - 'forest-green': Natural green palette\\n            - 'midnight-dark': Dark mode with bright accents\\n            - 'pastel-soft': Soft pastel colors\\n            - 'bold-contrast': High contrast bold colors\\n            - 'elegant-gold': Sophisticated gold and navy\\n            \\n            Select the theme that best matches the presentation's content and purpose.\\n            \",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"presentation_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the presentation (used for file naming)\"\n          },\n          \"title\": {\n            \"type\": \"string\",\n            \"description\": \"Main title of the presentation\"\n          },\n          \"subtitle\": {\n            \"type\": \"string\",\n            \"description\": \"Optional subtitle or tagline\"\n          },\n          \"theme\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"corporate-blue\",\n              \"modern-purple\",\n              \"minimal-mono\",\n              \"ocean-teal\",\n              \"sunset-warm\",\n              \"forest-green\",\n              \"midnight-dark\",\n              \"pastel-soft\",\n              \"bold-contrast\",\n              \"elegant-gold\"\n            ],\n            \"description\": \"Visual theme - choose based on content and purpose\"\n          },\n          \"slides\": {\n            \"type\": \"array\",\n            \"description\": \"Array of slide objects with layout and content\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"layout\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"title\",\n                    \"title-bullets\",\n                    \"title-content\",\n                    \"two-column\",\n                    \"image-text\",\n                    \"quote\",\n                    \"section\",\n                    \"blank\",\n                    \"hero-image\",\n                    \"image-grid\",\n                    \"comparison\",\n                    \"timeline\",\n                    \"stats\"\n                  ],\n                  \"description\": \"Layout type for the slide\"\n                },\n                \"content\": {\n                  \"type\": \"object\",\n                  \"description\": \"Content object specific to the layout type\",\n                  \"properties\": {\n                    \"title\": {\n                      \"type\": \"string\"\n                    },\n                    \"subtitle\": {\n                      \"type\": \"string\"\n                    },\n                    \"bullets\": {\n                      \"type\": \"array\",\n                      \"items\": {\n                        \"type\": \"string\"\n                      }\n                    },\n                    \"text\": {\n                      \"type\": \"string\"\n                    },\n                    \"left_content\": {\n                      \"type\": \"object\"\n                    },\n                    \"right_content\": {\n                      \"type\": \"object\"\n                    },\n                    \"image\": {\n                      \"type\": \"object\",\n                      \"properties\": {\n                        \"url\": {\n                          \"type\": \"string\"\n                        },\n                        \"alt\": {\n                          \"type\": \"string\"\n                        },\n                        \"position\": {\n                          \"type\": \"string\",\n                          \"enum\": [\n                            \"left\",\n                            \"right\",\n                            \"center\",\n                            \"background\"\n                          ]\n                        }\n                      }\n                    },\n                    \"quote\": {\n                      \"type\": \"string\"\n                    },\n                    \"author\": {\n                      \"type\": \"string\"\n                    },\n                    \"notes\": {\n                      \"type\": \"string\"\n                    }\n                  }\n                }\n              },\n              \"required\": [\n                \"layout\",\n                \"content\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"presentation_name\",\n          \"title\",\n          \"slides\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"export_presentation\",\n      \"description\": \"Export a JSON presentation to PPTX format with professional formatting.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"presentation_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the presentation to export\"\n          },\n          \"format\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"pptx\"\n            ],\n            \"description\": \"Export format (currently only PPTX supported)\"\n          }\n        },\n        \"required\": [\n          \"presentation_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"analyze_sheet\",\n      \"description\": \"Simple statistics and optional group_by; can export CSV.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"target_columns\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"group_by\": {\n            \"type\": \"string\"\n          },\n          \"aggregations\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\",\n              \"enum\": [\n                \"count\",\n                \"sum\",\n                \"avg\",\n                \"min\",\n                \"max\"\n              ]\n            }\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_sheet\",\n      \"description\": \"Create a new CSV/XLSX with optional headers/rows.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"headers\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"rows\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"array\",\n              \"items\": {\n                \"type\": \"string\"\n              }\n            }\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"overwrite\": {\n            \"type\": \"boolean\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"format_sheet\",\n      \"description\": \"Style and format sheet cells (bold headers, auto widths, optional conditional formatting). XLSX only.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"bold_headers\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"auto_width\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"apply_banding\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"conditional_format\": {\n            \"type\": \"object\",\n            \"properties\": {\n              \"column\": {\n                \"type\": \"string\"\n              },\n              \"min_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFEFEB\"\n              },\n              \"mid_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFD7D2\"\n              },\n              \"max_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFA39E\"\n              }\n            }\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_sheet\",\n      \"description\": \"Modify existing cells, rows, or columns (insert/delete/update).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"operations\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"type\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"update_cell\",\n                    \"update_row\",\n                    \"insert_row\",\n                    \"delete_row\",\n                    \"insert_column\",\n                    \"delete_column\"\n                  ]\n                },\n                \"row_index\": {\n                  \"type\": \"integer\"\n                },\n                \"column\": {\n                  \"type\": \"string\"\n                },\n                \"column_index\": {\n                  \"type\": \"integer\"\n                },\n                \"values\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  }\n                },\n                \"value\": {\n                  \"type\": \"string\"\n                }\n              },\n              \"required\": [\n                \"type\"\n              ]\n            }\n          },\n          \"save_as\": {\n            \"type\": \"string\",\n            \"description\": \"Optional path to save result (.csv or .xlsx)\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"operations\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"view_sheet\",\n      \"description\": \"Read headers, types, and sample rows; optional CSV export.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"max_rows\": {\n            \"type\": \"integer\",\n            \"default\": 100\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"visualize_sheet\",\n      \"description\": \"Generate charts (bar, line, pie, scatter) and save to XLSX. Also optionally export chart data to CSV for Google Sheets.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"chart_type\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"bar\",\n              \"line\",\n              \"pie\",\n              \"scatter\"\n            ],\n            \"default\": \"bar\"\n          },\n          \"x_column\": {\n            \"type\": \"string\"\n          },\n          \"y_columns\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"save_as\": {\n            \"type\": \"string\"\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\",\n            \"description\": \"Optional CSV path for the chart dataset (x + y columns)\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"x_column\",\n          \"y_columns\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"build_project\",\n      \"description\": \"Run production build for the project.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"project_name\": {\n            \"type\": \"string\"\n          },\n          \"package_manager\": {\n            \"type\": \"string\",\n            \"default\": \"pnpm\"\n          }\n        },\n        \"required\": [\n          \"project_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_project_structure\",\n      \"description\": \"Get the file structure of a web project, showing important files and directories.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"project_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the project directory to examine\"\n          },\n          \"max_depth\": {\n            \"type\": \"integer\",\n            \"description\": \"Maximum depth to traverse (default: 3)\",\n            \"default\": 3\n          }\n        },\n        \"required\": [\n          \"project_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"execute_data_provider_call\",\n      \"description\": \"Execute a call to a specific data provider endpoint\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"service_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the API service (e.g., 'linkedin')\"\n          },\n          \"route\": {\n            \"type\": \"string\",\n            \"description\": \"The key of the endpoint to call\"\n          },\n          \"payload\": {\n            \"type\": \"object\",\n            \"description\": \"The payload to send with the API call\"\n          }\n        },\n        \"required\": [\n          \"service_name\",\n          \"route\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_data_provider_endpoints\",\n      \"description\": \"Get available endpoints for a specific data provider\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"service_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the data provider (e.g., 'linkedin', 'twitter', 'zillow', 'amazon', 'yahoo_finance')\"\n          }\n        },\n        \"required\": [\n          \"service_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_current_agent_config\",\n      \"description\": \"Get the current configuration of the agent being edited. Use this to check what's already configured before making updates.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_agent\",\n      \"description\": \"Update the agent's configuration including name, description, system prompt, tools, and MCP servers. Call this whenever the user wants to modify any aspect of the agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the agent. Should be descriptive and indicate the agent's purpose.\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"A brief description of what the agent does and its capabilities.\"\n          },\n          \"system_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"The system instructions that define the agent's behavior, expertise, and approach. This should be comprehensive and well-structured.\"\n          },\n          \"agentpress_tools\": {\n            \"type\": \"object\",\n            \"description\": \"Configuration for AgentPress tools. Each key is a tool name, and the value is a boolean indicating if the tool is enabled.\",\n            \"additionalProperties\": {\n              \"type\": \"boolean\"\n            }\n          },\n          \"configured_mcps\": {\n            \"type\": \"array\",\n            \"description\": \"List of configured MCP servers for external integrations.\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"name\": {\n                  \"type\": \"string\"\n                },\n                \"qualifiedName\": {\n                  \"type\": \"string\"\n                },\n                \"config\": {\n                  \"type\": \"object\"\n                },\n                \"enabledTools\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  }\n                }\n              }\n            }\n          },\n          \"avatar\": {\n            \"type\": \"string\",\n            \"description\": \"Emoji to use as the agent's avatar.\"\n          },\n          \"avatar_color\": {\n            \"type\": \"string\",\n            \"description\": \"Hex color code for the agent's avatar background.\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"discover_user_mcp_servers\",\n      \"description\": \"Discover available MCP tools for a specific Composio profile. Use this to see what MCP tools are available for a connected profile.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The profile ID from the Composio credential profile\"\n          }\n        },\n        \"required\": [\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_app_details\",\n      \"description\": \"Get detailed information about a specific Composio toolkit, including available tools and authentication requirements.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"The toolkit slug to get details for (e.g., 'github', 'linear', 'slack')\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"search_mcp_servers\",\n      \"description\": \"Search for Composio toolkits based on user requirements. Use this when the user wants to add MCP tools to their agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"query\": {\n            \"type\": \"string\",\n            \"description\": \"Search query for finding relevant Composio toolkits (e.g., 'linear', 'github', 'database', 'search')\"\n          },\n          \"limit\": {\n            \"type\": \"integer\",\n            \"description\": \"Maximum number of toolkits to return (default: 10)\",\n            \"default\": 10\n          }\n        },\n        \"required\": [\n          \"query\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"configure_profile_for_agent\",\n      \"description\": \"Configure a connected credential profile to be used by the agent with selected tools. Use this after the profile is connected and you want to add it to the agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the connected credential profile\"\n          },\n          \"enabled_tools\": {\n            \"type\": \"array\",\n            \"description\": \"List of tool names to enable for this profile\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"display_name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional custom display name for this configuration in the agent\"\n          }\n        },\n        \"required\": [\n          \"profile_id\",\n          \"enabled_tools\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_credential_profile\",\n      \"description\": \"Create a new Composio credential profile for a specific toolkit. This will create the integration and return an authentication link that the user needs to visit to connect their account.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"The toolkit slug to create the profile for (e.g., 'github', 'linear', 'slack')\"\n          },\n          \"profile_name\": {\n            \"type\": \"string\",\n            \"description\": \"A name for this credential profile (e.g., 'Personal GitHub', 'Work Slack')\"\n          },\n          \"display_name\": {\n            \"type\": \"string\",\n            \"description\": \"Display name for the profile (defaults to profile_name if not provided)\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\",\n          \"profile_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_credential_profile\",\n      \"description\": \"Delete a credential profile that is no longer needed. This will also remove it from any agent configurations.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the credential profile to delete\"\n          }\n        },\n        \"required\": [\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_credential_profiles\",\n      \"description\": \"Get all existing Composio credential profiles for the current user. Use this to show the user their available profiles.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Optional filter to show only profiles for a specific toolkit\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"activate_workflow\",\n      \"description\": \"Activate or deactivate a workflow/playbook. Only active workflows can be executed.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to activate/deactivate\"\n          },\n          \"active\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to activate (true) or deactivate (false) the workflow\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_workflow\",\n      \"description\": \"Create a new workflow/playbook. This stores a Start node with a single child that contains config.playbook { template, variables }.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the workflow\"\n          },\n          \"template\": {\n            \"type\": \"string\",\n            \"description\": \"Workflow/playbook instructions text. Use {{variable}} tokens.\"\n          },\n          \"variables\": {\n            \"type\": \"array\",\n            \"description\": \"Optional variable specs; all treated as strings currently\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"key\": {\n                  \"type\": \"string\"\n                },\n                \"label\": {\n                  \"type\": \"string\"\n                },\n                \"required\": {\n                  \"type\": \"boolean\",\n                  \"default\": true\n                }\n              },\n              \"required\": [\n                \"key\"\n              ]\n            },\n            \"default\": []\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"Optional short description (auto-summarized from template if omitted)\"\n          },\n          \"is_default\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether this should be the default workflow\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"template\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_workflow\",\n      \"description\": \"Delete a workflow/playbook from the agent. This action cannot be undone.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to delete\"\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_workflows\",\n      \"description\": \"Get all workflows/playbooks for the current agent. Use this to see what workflows are already configured.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"include_steps\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include detailed step information for each workflow\",\n            \"default\": true\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_workflow\",\n      \"description\": \"Update an existing workflow or playbook. For playbooks, the first child under Start should contain config.playbook { template, variables }.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to update\"\n          },\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"New name for the workflow\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"New description for the workflow\"\n          },\n          \"trigger_phrase\": {\n            \"type\": \"string\",\n            \"description\": \"New trigger phrase for the workflow\"\n          },\n          \"is_default\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether this workflow should be the default workflow\"\n          },\n          \"status\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"draft\",\n              \"active\",\n              \"inactive\"\n            ],\n            \"description\": \"Status of the workflow\"\n          },\n          \"validate_tools\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to validate tool names against available tools when updating steps\",\n            \"default\": true\n          },\n          \"steps\": {\n            \"type\": \"array\",\n            \"description\": \"New steps. For playbooks, follow the Start + Execute Playbook structure with config.playbook.\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"name\": {\n                  \"type\": \"string\"\n                },\n                \"description\": {\n                  \"type\": \"string\"\n                },\n                \"type\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"instruction\",\n                    \"tool\",\n                    \"condition\"\n                  ],\n                  \"default\": \"instruction\"\n                },\n                \"config\": {\n                  \"type\": \"object\",\n                  \"additionalProperties\": true\n                },\n                \"conditions\": {\n                  \"type\": \"object\",\n                  \"additionalProperties\": true\n                },\n                \"order\": {\n                  \"type\": \"integer\"\n                },\n                \"children\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"$ref\": \"#\"\n                  }\n                }\n              },\n              \"required\": [\n                \"name\",\n                \"order\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_event_trigger\",\n      \"description\": \"Create a Composio event-based trigger for this agent. First list apps and triggers, then pass the chosen trigger slug, profile_id, and trigger_config. Optionally route to a workflow.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"slug\": {\n            \"type\": \"string\",\n            \"description\": \"Trigger type slug, e.g. 'GMAIL_NEW_GMAIL_MESSAGE'\"\n          },\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"Composio profile_id to use (must be connected)\"\n          },\n          \"trigger_config\": {\n            \"type\": \"object\",\n            \"description\": \"Trigger configuration object per trigger schema\",\n            \"additionalProperties\": true\n          },\n          \"route\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"agent\",\n              \"workflow\"\n            ],\n            \"default\": \"agent\",\n            \"description\": \"Execute agent directly or run a workflow\"\n          },\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional friendly name for the trigger\"\n          },\n          \"agent_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Prompt to pass to the agent when route is 'agent'\"\n          },\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"Workflow ID when route is 'workflow'\"\n          },\n          \"workflow_input\": {\n            \"type\": \"object\",\n            \"description\": \"Workflow input variables when route is 'workflow'\",\n            \"additionalProperties\": true\n          },\n          \"connected_account_id\": {\n            \"type\": \"string\",\n            \"description\": \"Connected account id; if omitted we try to derive from profile\"\n          }\n        },\n        \"required\": [\n          \"slug\",\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_scheduled_trigger\",\n      \"description\": \"Create a scheduled trigger for the agent to execute workflows or direct agent runs using cron expressions. This allows the agent to run automatically at specified times.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the scheduled trigger. Should be descriptive of when/why it runs.\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"Description of what this trigger does and when it runs.\"\n          },\n          \"cron_expression\": {\n            \"type\": \"string\",\n            \"description\": \"Cron expression defining when to run (e.g., '0 9 * * *' for daily at 9am, '*/30 * * * *' for every 30 minutes)\"\n          },\n          \"execution_type\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"workflow\",\n              \"agent\"\n            ],\n            \"description\": \"Whether to execute a workflow or run the agent directly\",\n            \"default\": \"agent\"\n          },\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to execute (required if execution_type is 'workflow')\"\n          },\n          \"workflow_input\": {\n            \"type\": \"object\",\n            \"description\": \"Input data to pass to the workflow (optional, only for workflow execution)\",\n            \"additionalProperties\": true\n          },\n          \"agent_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Prompt to send to the agent when triggered (required if execution_type is 'agent')\"\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"cron_expression\",\n          \"execution_type\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_scheduled_trigger\",\n      \"description\": \"Delete a scheduled trigger. The agent will no longer run automatically at the scheduled time.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"trigger_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the trigger to delete\"\n          }\n        },\n        \"required\": [\n          \"trigger_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_scheduled_triggers\",\n      \"description\": \"Get all scheduled triggers for the current agent. Shows when the agent will run automatically.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_app_event_triggers\",\n      \"description\": \"List available triggers for a given app/toolkit slug. Includes slug, name, description, type, instructions, config, and payload schema.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Toolkit slug, e.g. 'gmail'\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_event_profiles\",\n      \"description\": \"List connected Composio profiles for a toolkit. Use this to get profile_id and connected_account_id before creating a trigger.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Toolkit slug, e.g. 'gmail'\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_event_trigger_apps\",\n      \"description\": \"List apps (toolkits) that have available event-based triggers via Composio. Returns slug, name, and logo.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"toggle_scheduled_trigger\",\n      \"description\": \"Enable or disable a scheduled trigger. Disabled triggers won't run until re-enabled.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"trigger_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the trigger to toggle\"\n          },\n          \"is_active\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to enable (true) or disable (false) the trigger\"\n          }\n        },\n        \"required\": [\n          \"trigger_id\",\n          \"is_active\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_act\",\n      \"description\": \"Perform any browser action using natural language description. CRITICAL: This tool automatically provides a screenshot with every action. For data entry actions (filling forms, entering text, selecting options), you MUST review the provided screenshot to verify that displayed values exactly match what was intended. Report mismatches immediately.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"action\": {\n            \"type\": \"string\",\n            \"description\": \"The action to perform. Examples: 'click the login button', 'fill in the email field with test@example.com', 'scroll down to see more content', 'select option 2 from the dropdown', 'press Enter', 'go back', 'wait 5 seconds', 'click at coordinates 100,200', 'drag the file icon to the drop zone', 'select United States from the country dropdown'\"\n          },\n          \"variables\": {\n            \"type\": \"object\",\n            \"description\": \"Variables to use in the action. Variables in the action string are referenced using %variable_name%. These variables are NOT shared with LLM providers for security.\",\n            \"additionalProperties\": {\n              \"type\": \"string\"\n            },\n            \"default\": {}\n          },\n          \"iframes\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include iframe content in the action. Set to true if the target element is inside an iframe.\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"action\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_extract_content\",\n      \"description\": \"Extract structured content from the current page using Stagehand\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"instruction\": {\n            \"type\": \"string\",\n            \"description\": \"What content to extract (e.g., 'extract all product prices', 'get the main heading', 'extract apartment listings with address and price')\"\n          },\n          \"selector\": {\n            \"type\": \"string\",\n            \"description\": \"Optional XPath selector to reduce extraction scope to a specific element. Useful for reducing input tokens and increasing accuracy.\",\n            \"default\": null\n          },\n          \"iframes\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include iframe content in the extraction. Set to true if the target content is inside an iframe.\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"instruction\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_navigate_to\",\n      \"description\": \"Navigate to a specific url\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"url\": {\n            \"type\": \"string\",\n            \"description\": \"The url to navigate to\"\n          }\n        },\n        \"required\": [\n          \"url\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_screenshot\",\n      \"description\": \"Take a screenshot of the current page\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the screenshot\",\n            \"default\": \"screenshot\"\n          }\n        }\n      }\n    }\n  }\n]\n```\n\nWhen using the tools:\n- Use the exact function names from the JSON schema above\n- Include all required parameters as specified in the schema\n- Format complex data (objects, arrays) as JSON strings within the parameter tags\n- Boolean values should be \"true\" or \"false\" (lowercase)\n\n\nUsage Examples:\n\nexpand_message:\n\n        <!-- Example 1: Expand a message that was truncated in the previous conversation -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">ecde3a4c-c7dc-4776-ae5c-8209517c5576</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Expand a message to create reports or analyze truncated data -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">f47ac10b-58cc-4372-a567-0e02b2c3d479</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Expand a message when you need the full content for analysis -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">550e8400-e29b-41d4-a716-446655440000</parameter>\n        </invoke>\n        </function_calls>\n        \n\nask:\n\n        <function_calls>\n        <invoke name=\"ask\">\n        <parameter name=\"text\">I'm planning to bake the chocolate cake for your birthday party. The recipe mentions \"rich frosting\" but doesn't specify what type. Could you clarify your preferences? For example:\n1. Would you prefer buttercream or cream cheese frosting?\n2. Do you want any specific flavor added to the frosting (vanilla, coffee, etc.)?\n3. Should I add any decorative toppings like sprinkles or fruit?\n4. Do you have any dietary restrictions I should be aware of?\n\nThis information will help me make sure the cake meets your expectations for the celebration.</parameter>\n        <parameter name=\"attachments\">recipes/chocolate_cake.txt,photos/cake_examples.jpg</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncomplete:\n\n        <function_calls>\n        <invoke name=\"complete\">\n        <parameter name=\"text\">I have successfully completed all tasks for your project. Here's what was accomplished:\n1. Created the web application with modern UI components\n2. Implemented user authentication and database integration\n3. Deployed the application to production\n4. Created comprehensive documentation\n\nAll deliverables are attached for your review.</parameter>\n        <parameter name=\"attachments\">app/src/main.js,docs/README.md,deployment-config.yaml</parameter>\n        </invoke>\n        </function_calls>\n        \n\nweb_browser_takeover:\n\n        <function_calls>\n        <invoke name=\"web_browser_takeover\">\n        <parameter name=\"text\">I've encountered a CAPTCHA verification on the page that I can't solve automatically. Could you help me out?\n\nHere's what I need you to do:\n1. Solve the CAPTCHA puzzle that's currently displayed\n2. Let me know once you've completed it\n3. I'll then continue with the automated process\n\nIf you encounter any issues or need to take additional steps, please let me know. Thanks for your help!</parameter>\n        </invoke>\n        </function_calls>\n        \n\nclear_all:\n\n        <function_calls>\n        <invoke name=\"clear_all\">\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_tasks:\n\n        # Batch creation across multiple sections:\n        <function_calls>\n        <invoke name=\"create_tasks\">\n        <parameter name=\"sections\">[\n            {\n                \"title\": \"Setup & Planning\", \n                \"tasks\": [\"Research requirements\", \"Create project plan\"]\n            },\n            {\n                \"title\": \"Development\", \n                \"tasks\": [\"Setup environment\", \"Write code\", \"Add tests\"]\n            },\n            {\n                \"title\": \"Deployment\", \n                \"tasks\": [\"Deploy to staging\", \"Run tests\", \"Deploy to production\"]\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Simple single section creation:\n        <function_calls>\n        <invoke name=\"create_tasks\">\n        <parameter name=\"section_title\">Bug Fixes</parameter>\n        <parameter name=\"task_contents\">[\"Fix login issue\", \"Update error handling\"]</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_tasks:\n\n        # Delete single task:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">task-uuid-here</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete multiple tasks:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\"]</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete single section (and all its tasks):\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"section_ids\">section-uuid-here</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete multiple sections (and all their tasks):\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"section_ids\">[\"section-id-1\", \"section-id-2\"]</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete both tasks and sections:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\"]</parameter>\n        <parameter name=\"section_ids\">[\"section-id-1\"]</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nupdate_tasks:\n\n        # Update single task (when only one task is completed):\n        <function_calls>\n        <invoke name=\"update_tasks\">\n        <parameter name=\"task_ids\">task-uuid-here</parameter>\n        <parameter name=\"status\">completed</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Update multiple tasks (EFFICIENT: batch multiple completed tasks):\n        <function_calls>\n        <invoke name=\"update_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\", \"task-id-3\"]</parameter>\n        <parameter name=\"status\">completed</parameter>\n        </invoke>\n        </function_calls>\n        \n\nview_tasks:\n\n        <function_calls>\n        <invoke name=\"view_tasks\">\n        </invoke>\n        </function_calls>\n        \n\ncheck_command_output:\n\n        <function_calls>\n        <invoke name=\"check_command_output\">\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n        \n        <!-- Example 2: Check final output and kill session -->\n        <function_calls>\n        <invoke name=\"check_command_output\">\n        <parameter name=\"session_name\">build_process</parameter>\n        <parameter name=\"kill_session\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexecute_command:\n\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm run dev</parameter>\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Running in Specific Directory -->\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm run build</parameter>\n        <parameter name=\"folder\">frontend</parameter>\n        <parameter name=\"session_name\">build_process</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Blocking command (wait for completion) -->\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm install</parameter>\n        <parameter name=\"blocking\">true</parameter>\n        <parameter name=\"timeout\">300</parameter>\n        </invoke>\n        </function_calls>\n        \n\nlist_commands:\n\n        <function_calls>\n        <invoke name=\"list_commands\">\n        </invoke>\n        </function_calls>\n        \n\nterminate_command:\n\n        <function_calls>\n        <invoke name=\"terminate_command\">\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_file:\n\n        <function_calls>\n        <invoke name=\"create_file\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"file_contents\">\n        # This is the file content\n        def main():\n            print(\"Hello, World!\")\n        \n        if __name__ == \"__main__\":\n            main()\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_file:\n\n        <function_calls>\n        <invoke name=\"delete_file\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        </invoke>\n        </function_calls>\n        \n\nedit_file:\n\n        <!-- Example: Mark multiple scattered tasks as complete in a todo list -->\n        <function_calls>\n        <invoke name=\"edit_file\">\n        <parameter name=\"target_file\">todo.md</parameter>\n        <parameter name=\"instructions\">I am marking the research and setup tasks as complete in my todo list.</parameter>\n        <parameter name=\"code_edit\">\n// ... existing code ...\n- [x] Research topic A\n- [ ] Research topic B\n- [x] Research topic C\n// ... existing code ...\n- [x] Setup database\n- [x] Configure server\n// ... existing code ...\n        </parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example: Add error handling and logging to a function -->\n        <function_calls>\n        <invoke name=\"edit_file\">\n        <parameter name=\"target_file\">src/main.py</parameter>\n        <parameter name=\"instructions\">I am adding error handling and logging to the user authentication function</parameter>\n        <parameter name=\"code_edit\">\n// ... existing imports ...\nfrom my_app.logging import logger\nfrom my_app.exceptions import DatabaseError\n// ... existing code ...\ndef authenticate_user(username, password):\n    try:\n        user = get_user(username)\n        if user and verify_password(password, user.password_hash):\n            return user\n        return None\n    except DatabaseError as e:\n        logger.error(f\"Database error during authentication: {e}\")\n        return None\n    except Exception as e:\n        logger.error(f\"Unexpected error during authentication: {e}\")\n        return None\n// ... existing code ...\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\nfull_file_rewrite:\n\n        <function_calls>\n        <invoke name=\"full_file_rewrite\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"file_contents\">\n        This completely replaces the entire file content.\n        Use when making major changes to a file or when the changes\n        are too extensive for str-replace.\n        All previous content will be lost and replaced with this text.\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\nstr_replace:\n\n        <function_calls>\n        <invoke name=\"str_replace\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"old_str\">text to replace (must appear exactly once in the file)</parameter>\n        <parameter name=\"new_str\">replacement text that will be inserted instead</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndeploy:\n\n        <!-- \n        IMPORTANT: Only use this tool when:\n        1. The user explicitly requests permanent deployment to production\n        2. You have a complete, ready-to-deploy directory \n        \n        NOTE: If the same name is used, it will redeploy to the same project as before\n        -->\n\n        <function_calls>\n        <invoke name=\"deploy\">\n        <parameter name=\"name\">my-site</parameter>\n        <parameter name=\"directory_path\">website</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexpose_port:\n\n        <!-- Example 1: Expose a web server running on port 8000 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">8000</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Expose an API service running on port 3000 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">3000</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Expose a development server running on port 5173 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">5173</parameter>\n        </invoke>\n        </function_calls>\n        \n\nscrape_webpage:\n\n        <function_calls>\n        <invoke name=\"scrape_webpage\">\n        <parameter name=\"urls\">https://www.kortix.ai/,https://github.com/kortix-ai/suna</parameter>\n        </invoke>\n        </function_calls>\n        \n\nweb_search:\n\n        <function_calls>\n        <invoke name=\"web_search\">\n        <parameter name=\"query\">what is Kortix AI and what are they building?</parameter>\n        <parameter name=\"num_results\">20</parameter>\n        </invoke>\n        </function_calls>\n        \n        <!-- Another search example -->\n        <function_calls>\n        <invoke name=\"web_search\">\n        <parameter name=\"query\">latest AI research on transformer models</parameter>\n        <parameter name=\"num_results\">20</parameter>\n        </invoke>\n        </function_calls>\n        \n\nsee_image:\n\n        <!-- Example: Request to see a local image named 'diagram.png' inside the 'docs' folder -->\n        <function_calls>\n        <invoke name=\"see_image\">\n        <parameter name=\"file_path\">docs/diagram.png</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example: Request to see an image from a URL -->\n        <function_calls>\n        <invoke name=\"see_image\">\n        <parameter name=\"file_path\">https://example.com/image.jpg</parameter>\n        </invoke>\n        </function_calls>\n        \n\nimage_edit_or_generate:\n\n        Generate mode example (new image):\n        <function_calls>\n        <invoke name=\"image_edit_or_generate\">\n        <parameter name=\"mode\">generate</parameter>\n        <parameter name=\"prompt\">A futuristic cityscape at sunset</parameter>\n        </invoke>\n        </function_calls>\n        \n        Edit mode example (modifying existing):\n        <function_calls>\n        <invoke name=\"image_edit_or_generate\">\n        <parameter name=\"mode\">edit</parameter>\n        <parameter name=\"prompt\">Add a red hat to the person in the image</parameter>\n        <parameter name=\"image_path\">generated_image_abc123.png</parameter>\n        </invoke>\n        </function_calls>\n        \n        Multi-turn workflow (follow-up edits):\n        1. User: \"Create a logo\" \u2192 generate mode\n        2. User: \"Make it more colorful\" \u2192 edit mode (automatic)\n        3. User: \"Add text to it\" \u2192 edit mode (automatic)\n        \n\ncreate_presentation_outline:\n\n        <function_calls>\n        <invoke name=\"create_presentation_outline\">\n        <parameter name=\"title\">The Future of AI</parameter>\n        <parameter name=\"subtitle\">Transforming how we work, create, and connect</parameter>\n        <parameter name=\"slides\">[\n            {\n                \"title\": \"The Future of AI\",\n                \"description\": \"Title slide with bold typography and clean flat design, setting a professional tone for the presentation.\",\n                \"notes\": \"Open with confidence. Use a solid color background with high contrast text. Keep the design minimal and impactful.\"\n            },\n            {\n                \"title\": \"Revolutionary Technology\",\n                \"description\": \"Content slide showcasing key AI capabilities with clean typography and simple bullet points.\",\n                \"notes\": \"Highlight the four key areas: neural networks, automation, creativity, and connectivity. Use flat icons or simple graphics.\"\n            },\n            {\n                \"title\": \"The Question\",\n                \"description\": \"Minimal slide with large, bold text asking a thought-provoking question.\",\n                \"notes\": \"Pause for emphasis. Use a contrasting flat color background. Let the question sink in before continuing.\"\n            },\n            {\n                \"title\": \"Real-World Impact\",\n                \"description\": \"Showcase concrete AI applications across industries with clean layout and organized content sections.\",\n                \"notes\": \"Present examples from healthcare, transportation, climate, and education using flat design principles.\"\n            },\n            {\n                \"title\": \"The Road Ahead\",\n                \"description\": \"Inspirational slide with a powerful quote and forward-looking message about human potential.\",\n                \"notes\": \"Quote: 'The best way to predict the future is to invent it.' Use clean typography and solid colors.\"\n            },\n            {\n                \"title\": \"Thank You\",\n                \"description\": \"Clean, minimal closing slide inviting questions, using flat design with good typography.\",\n                \"notes\": \"Keep it simple. Use a solid background color with contrasting text. Create space for audience engagement.\"\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_presentation:\n\n        <function_calls>\n        <invoke name=\"create_presentation\">\n        <parameter name=\"presentation_name\">company_overview</parameter>\n        <parameter name=\"title\">Company Overview 2024</parameter>\n        <parameter name=\"subtitle\">Innovation Through Technology</parameter>\n        <parameter name=\"theme\">modern-purple</parameter>\n        <parameter name=\"slides\">[\n            {\n                \"layout\": \"hero-image\",\n                \"content\": {\n                    \"title\": \"Welcome to the Future\",\n                    \"subtitle\": \"Where Innovation Meets Excellence\",\n                    \"image\": {\n                        \"url\": \"unsplash:technology office\",\n                        \"position\": \"background\"\n                    }\n                }\n            },\n            {\n                \"layout\": \"stats\",\n                \"content\": {\n                    \"title\": \"Our Impact in Numbers\",\n                    \"stats\": [\n                        {\"value\": \"50K+\", \"label\": \"Active Users\"},\n                        {\"value\": \"$2.5M\", \"label\": \"Revenue\"},\n                        {\"value\": \"98%\", \"label\": \"Satisfaction\"},\n                        {\"value\": \"15\", \"label\": \"Countries\"}\n                    ]\n                }\n            },\n            {\n                \"layout\": \"image-text\",\n                \"content\": {\n                    \"title\": \"Our Mission\",\n                    \"text\": \"We empower businesses with cutting-edge AI technology to transform their operations, enhance productivity, and unlock new possibilities. Our platform combines powerful automation with human creativity.\",\n                    \"image\": {\n                        \"url\": \"unsplash:artificial intelligence\",\n                        \"position\": \"right\"\n                    }\n                }\n            },\n            {\n                \"layout\": \"image-grid\",\n                \"content\": {\n                    \"title\": \"Our Products in Action\",\n                    \"images\": [\n                        {\"url\": \"unsplash:dashboard analytics\", \"caption\": \"Real-time Analytics\"},\n                        {\"url\": \"unsplash:team collaboration\", \"caption\": \"Team Collaboration\"},\n                        {\"url\": \"unsplash:mobile app\", \"caption\": \"Mobile Experience\"},\n                        {\"url\": \"unsplash:data visualization\", \"caption\": \"Data Insights\"}\n                    ]\n                }\n            },\n            {\n                \"layout\": \"title-bullets\",\n                \"content\": {\n                    \"title\": \"Key Features\",\n                    \"bullets\": [\n                        \"AI-powered automation for repetitive tasks\",\n                        \"Real-time collaboration and communication\",\n                        \"Advanced analytics and reporting\",\n                        \"Enterprise-grade security and compliance\",\n                        \"24/7 customer support and training\"\n                    ]\n                }\n            },\n            {\n                \"layout\": \"quote\",\n                \"content\": {\n                    \"quote\": \"This platform has transformed how we work. We've saved 40% of our time and increased productivity by 60%.\",\n                    \"author\": \"Sarah Johnson, CTO at TechCorp\"\n                }\n            },\n            {\n                \"layout\": \"section\",\n                \"content\": {\n                    \"title\": \"Let's Build Together\",\n                    \"subtitle\": \"Start Your Journey Today\"\n                }\n            },\n            {\n                \"layout\": \"title\",\n                \"content\": {\n                    \"title\": \"Thank You\",\n                    \"subtitle\": \"Questions? Let's Connect!\",\n                    \"notes\": \"Contact us at hello@company.com\"\n                }\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\nanalyze_sheet:\n\n        <function_calls>\n        <invoke name=\"analyze_sheet\">\n          <parameter name=\"file_path\">data/sales.csv</parameter>\n          <parameter name=\"target_columns\">[\"revenue\"]</parameter>\n          <parameter name=\"group_by\">region</parameter>\n          <parameter name=\"export_csv_path\">data/sales_summary.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_sheet:\n\n        <function_calls>\n        <invoke name=\"create_sheet\">\n          <parameter name=\"file_path\">data/sample.csv</parameter>\n          <parameter name=\"headers\">[\"region\",\"revenue\"]</parameter>\n          <parameter name=\"rows\">[[\"NA\",100],[\"EU\",120]]</parameter>\n          <parameter name=\"overwrite\">true</parameter>\n        </invoke>\n        </function_calls>\n    \n\nformat_sheet:\n\n        <function_calls>\n        <invoke name=\"format_sheet\">\n          <parameter name=\"file_path\">reports/sales_chart.xlsx</parameter>\n          <parameter name=\"sheet_name\">Data</parameter>\n          <parameter name=\"bold_headers\">true</parameter>\n          <parameter name=\"conditional_format\">{\"column\":\"revenue\"}</parameter>\n        </invoke>\n        </function_calls>\n    \n\nupdate_sheet:\n\n        <function_calls>\n        <invoke name=\"update_sheet\">\n          <parameter name=\"file_path\">employee_details.xlsx</parameter>\n          <parameter name=\"sheet_name\">Employee Database</parameter>\n          <parameter name=\"operations\">[\n            {\"type\":\"insert_row\",\"row_index\":6,\"values\":[\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]},\n            {\"type\":\"insert_row\",\"row_index\":7,\"values\":[\"SUMMARY CALCULATIONS\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]},\n            {\"type\":\"insert_row\",\"row_index\":8,\"values\":[\"Total Employees:\",\"=COUNTA(A2:A100)\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]}\n          ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\nview_sheet:\n\n        <function_calls>\n        <invoke name=\"view_sheet\">\n          <parameter name=\"file_path\">reports/sales.xlsx</parameter>\n          <parameter name=\"sheet_name\">Sheet1</parameter>\n          <parameter name=\"max_rows\">50</parameter>\n          <parameter name=\"export_csv_path\">reports/sales_preview.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\nvisualize_sheet:\n\n        <function_calls>\n        <invoke name=\"visualize_sheet\">\n          <parameter name=\"file_path\">reports/sales.csv</parameter>\n          <parameter name=\"x_column\">region</parameter>\n          <parameter name=\"y_columns\">[\"revenue\"]</parameter>\n          <parameter name=\"chart_type\">bar</parameter>\n          <parameter name=\"save_as\">reports/sales_chart.xlsx</parameter>\n          <parameter name=\"export_csv_path\">reports/sales_chart_data.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\nbuild_project:\n\n        <!-- Build a Next.js project for production -->\n        <function_calls>\n        <invoke name=\"build_project\">\n        <parameter name=\"project_name\">my-app</parameter>\n        <parameter name=\"package_manager\">pnpm</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_project_structure:\n\n        <!-- Get structure of a project -->\n        <function_calls>\n        <invoke name=\"get_project_structure\">\n        <parameter name=\"project_name\">my-app</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexecute_data_provider_call:\n\n        <!-- \n        The execute-data-provider-call tool makes a request to a specific data provider endpoint.\n        Use this tool when you need to call an data provider endpoint with specific parameters.\n        The route must be a valid endpoint key obtained from get-data-provider-endpoints tool!!\n        -->\n        \n        <!-- Example to call linkedIn service with the specific route person -->\n        <function_calls>\n        <invoke name=\"execute_data_provider_call\">\n        <parameter name=\"service_name\">linkedin</parameter>\n        <parameter name=\"route\">person</parameter>\n        <parameter name=\"payload\">{\"link\": \"https://www.linkedin.com/in/johndoe/\"}</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_data_provider_endpoints:\n\n<!-- \nThe get-data-provider-endpoints tool returns available endpoints for a specific data provider.\nUse this tool when you need to discover what endpoints are available.\n-->\n\n<!-- Example to get LinkedIn API endpoints -->\n<function_calls>\n<invoke name=\"get_data_provider_endpoints\">\n<parameter name=\"service_name\">linkedin</parameter>\n</invoke>\n</function_calls>\n        \n\nget_current_agent_config:\n\n        <function_calls>\n        <invoke name=\"get_current_agent_config\">\n        </invoke>\n        </function_calls>\n        \n\nupdate_agent:\n\n        <function_calls>\n        <invoke name=\"update_agent\">\n        <parameter name=\"name\">Research Assistant</parameter>\n        <parameter name=\"description\">An AI assistant specialized in conducting research and providing comprehensive analysis</parameter>\n        <parameter name=\"system_prompt\">You are a research assistant with expertise in gathering, analyzing, and synthesizing information. Your approach is thorough and methodical...</parameter>\n                        <parameter name=\"agentpress_tools\">{\"web_search_tool\": true, \"sb_files_tool\": true, \"sb_shell_tool\": false}</parameter>\n        <parameter name=\"avatar\">\ud83d\udd2c</parameter>\n        <parameter name=\"avatar_color\">#4F46E5</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndiscover_user_mcp_servers:\n\n        <function_calls>\n        <invoke name=\"discover_user_mcp_servers\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_app_details:\n\n        <function_calls>\n        <invoke name=\"get_app_details\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        </invoke>\n        </function_calls>\n        \n\nsearch_mcp_servers:\n\n        <function_calls>\n        <invoke name=\"search_mcp_servers\">\n        <parameter name=\"query\">linear</parameter>\n        <parameter name=\"limit\">5</parameter>\n        </invoke>\n        </function_calls>\n        \n\nconfigure_profile_for_agent:\n\n        <function_calls>\n        <invoke name=\"configure_profile_for_agent\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        <parameter name=\"enabled_tools\">[\"create_issue\", \"list_repositories\", \"get_user\"]</parameter>\n        <parameter name=\"display_name\">Personal GitHub Integration</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_credential_profile:\n\n        <function_calls>\n        <invoke name=\"create_credential_profile\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        <parameter name=\"profile_name\">Personal GitHub</parameter>\n        <parameter name=\"display_name\">My Personal GitHub Account</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_credential_profile:\n\n        <function_calls>\n        <invoke name=\"delete_credential_profile\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_credential_profiles:\n\n        <function_calls>\n        <invoke name=\"get_credential_profiles\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        </invoke>\n        </function_calls>\n        \n\nactivate_workflow:\n\n        <function_calls>\n        <invoke name=\"activate_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"active\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_workflow:\n\n        <function_calls>\n        <invoke name=\"create_workflow\">\n        <parameter name=\"name\">Company Research Playbook</parameter>\n        <parameter name=\"template\">Research companies and update sheets using {{sheet_id}} and {{limit}}</parameter>\n        <parameter name=\"variables\">[{\"key\":\"sheet_id\",\"label\":\"Sheet ID\",\"required\":true},{\"key\":\"limit\",\"label\":\"Row Limit\",\"required\":false}]</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_workflow:\n\n        <function_calls>\n        <invoke name=\"delete_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_workflows:\n\n        <function_calls>\n        <invoke name=\"get_workflows\">\n        <parameter name=\"include_steps\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nupdate_workflow:\n\n        <function_calls>\n        <invoke name=\"update_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"name\">Updated Research Workflow</parameter>\n        <parameter name=\"status\">active</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_event_trigger:\n\n        <function_calls>\n        <invoke name=\"list_event_trigger_apps\"></invoke>\n        <invoke name=\"list_app_event_triggers\"><parameter name=\"toolkit_slug\">gmail</parameter></invoke>\n        <invoke name=\"list_event_profiles\"><parameter name=\"toolkit_slug\">gmail</parameter></invoke>\n        <invoke name=\"create_event_trigger\">\n          <parameter name=\"slug\">GMAIL_NEW_GMAIL_MESSAGE</parameter>\n          <parameter name=\"profile_id\">profile_123</parameter>\n          <parameter name=\"trigger_config\">{\"interval\": 1, \"userId\": \"me\", \"labelIds\": \"INBOX\"}</parameter>\n          <parameter name=\"route\">agent</parameter>\n          <parameter name=\"agent_prompt\">Read this</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"create_scheduled_trigger\">\n        <parameter name=\"name\">Daily Report Generation</parameter>\n        <parameter name=\"description\">Generates daily reports every morning at 9 AM</parameter>\n        <parameter name=\"cron_expression\">0 9 * * *</parameter>\n        <parameter name=\"execution_type\">workflow</parameter>\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"workflow_input\">{\"report_type\": \"daily\", \"include_charts\": true}</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"delete_scheduled_trigger\">\n        <parameter name=\"trigger_id\">trigger-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_scheduled_triggers:\n\n        <function_calls>\n        <invoke name=\"get_scheduled_triggers\">\n        </invoke>\n        </function_calls>\n        \n\nlist_app_event_triggers:\n\n        <function_calls>\n        <invoke name=\"list_app_event_triggers\">\n        <parameter name=\"toolkit_slug\">gmail</parameter>\n        </invoke>\n        </function_calls>\n    \n\nlist_event_profiles:\n\n        <function_calls>\n        <invoke name=\"list_event_profiles\">\n        <parameter name=\"toolkit_slug\">gmail</parameter>\n        </invoke>\n        </function_calls>\n    \n\nlist_event_trigger_apps:\n\n        <function_calls>\n        <invoke name=\"list_event_trigger_apps\"></invoke>\n        </function_calls>\n    \n\ntoggle_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"toggle_scheduled_trigger\">\n        <parameter name=\"trigger_id\">trigger-123</parameter>\n        <parameter name=\"is_active\">false</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_act:\n\n        <function_calls>\n        <invoke name=\"browser_act\">\n        <parameter name=\"action\">fill in the login form with %username% and %password%</parameter>\n        <parameter name=\"variables\">{\"username\": \"john.doe\", \"password\": \"secret123\"}</parameter>\n        <parameter name=\"iframes\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_extract_content:\n\n        <function_calls>\n        <invoke name=\"browser_extract_content\">\n        <parameter name=\"instruction\">extract all product names and prices from the main product list</parameter>\n        <parameter name=\"selector\">//div[@class='product-list']</parameter>\n        <parameter name=\"iframes\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_navigate_to:\n\n        <function_calls>\n        <invoke name=\"browser_navigate_to\">\n        <parameter name=\"url\">https://example.com</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_screenshot:\n\n        <function_calls>\n        <invoke name=\"browser_screenshot\">\n        <parameter name=\"name\">page_screenshot</parameter>\n        </invoke>\n        </function_calls>\n        \n"
    },
    {
      "role": "user",
      "content": "1. Research {{company_name}} comprehensively\r\n2. Gather recent news, funding, and leadership info\r\n3. Analyze competitive position and market share\r\n4. Research key personnel background\r\n5. Create detailed profile with actionable insights",
      "message_id": "0e8ffd17-347d-47d8-9f4f-610dde716202"
    },
    {
      "role": "user",
      "content": "Do it",
      "message_id": "70cbfcc1-8da3-4af0-a729-32456015be1d"
    },
    {
      "role": "user",
      "content": "DO IT",
      "message_id": "04a8eed9-a56d-4165-b7dd-79086671c247"
    },
    {
      "role": "user",
      "content": "DO IT",
      "message_id": "88b8b0a7-59f8-49c4-aa60-a64af4f73246"
    },
    {
      "role": "assistant",
      "content": "I'll help you create a comprehensive research workflow for company analysis. Let me set up a structured approach to gather all the information you need.\n\n<function_calls>\n<invoke name=\"create_tasks\">\n<parameter name=\"sections\">[\n  {\n    \"title\": \"Research Setup & Planning\",\n    \"tasks\": [\n      \"Set up company research workflow with variables\",\n      \"Configure data sources and sea... (truncated)\n\nmessage_id \"9839eb81-97b8-46a5-84ec-b271f905b24e\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "9839eb81-97b8-46a5-84ec-b271f905b24e"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"create_tasks\", \"xml_tag_name\": \"create-tasks\", \"tool_call_id\": null, \"arguments\": {\"sections\": [{\"title\": \"Research Setup & Planning\", \"tasks\": [\"Set up company research workflow with variables\", \"Configure data sources and search parameters\", \"Establish research methodology and sources\"]}, {\"title\": \"Company Fundamentals Research\", \"tasks\": [\"Research basic company information and overview\", \"Gather recent news and announcements\", \"Research funding history and investment rounds\", \"Analyze financial performance and metrics\"]}, {\"title\": \"Leadership & Personnel Analysis\", \"tasks\": [\"Research CEO and executive team backgrounds\", \"Analyze key personnel LinkedIn profiles\", \"Research board members and advisors\", \"Identify leadership changes and transitions\"]}, {\"title\": \"Market Position Analysis\", \"tasks\": [\"Research competitive landscape\", \"Analyze market share and positioning\", \"Identify key competitors and differentiators\", \"Research industry trends and market dynamics\"]}, {\"title\": \"Insights & Profile Creation\", \"tasks\": [\"Synthesize all research findings\", \"Create comprehensive company profile\", \"Generate actionable insights and recommendations\", \"Format final deliverable with citations\"]}]}, \"result\": {\"success\": true, \"output\": {\"sections\": [{\"id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\", \"title\": \"Research Setup & Planning\", \"tasks\": [{\"id\": \"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\", \"content\": \"Set up company research workflow with variables\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"content\": \"Configure data sources and search parameters\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"20b5484e-fe54-4236-8f5c-e71a738a26da\", \"content\": \"Establish research methodology and sources\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}]}, {\"id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\", \"title\": \"Company Fundamentals Research\", \"tasks\": [{\"id\": \"b08bc6af-f98c-46a4-b965-8c96c963ea33\", \"content\": \"Research basic company information and overview\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"d97ef9f3-3d4a-47b0-8ae6-b8137c11ea32\", \"content\": \"Gather recent news and announcements\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"a93178e5-a6d7-4c2c-8a30-50bb78987d36\", \"content\": \"Research funding history and investment rounds\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"7ac8c5a2-8cc0-4cfa-9202-44812e7a6a33\", \"content\": \"Analyze financial performance and metrics\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}]}, {\"id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\", \"title\": \"Leadership & Personnel Analysis\", \"tasks\": [{\"id\": \"a1ae7fc4-5abd-4197-b783-319c874de228\", \"content\": \"Research CEO and executive team backgrounds\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"238940d8-d6cc-45e7-83b3-a6377137201c\", \"content\": \"Analyze key personnel LinkedIn profiles\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"ddbe449e-33a5-4675-862e-43775a0b2730\", \"content\": \"Research board members and advisors\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"643e917a-088c-4140-8426-cafb7e600faa\", \"content\": \"Identify leadership changes and transitions\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}]}, {\"id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\", \"title\": \"Market Position Analysis\", \"tasks\": [{\"id\": \"f4e90132-5b18-4b3c-9018-39c911414e9b\", \"content\": \"Research competitive landscape\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"454b1d2b-4da1-4c14-8f6e-31073a789702\", \"content\": \"Analyze market share and positioning\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"92282e56-bee1-44c2-a4ba-7679472ea5d1\", \"content\": \"Identify key competitors and differentiators\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"aa1fe7aa-c4ec-44c5-85d3-2cac70fe88c7\", \"content\": \"Research industry trends and market dynamics\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}]}, {\"id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\", \"title\": \"Insights & Profile Creation\", \"tasks\": [{\"id\": \"f78993ad-b1b9-4fc8-8896-5616c3da35c1\", \"content\": \"Synthesize all research findings\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"5e1421e0-55b6-4bdc-9d20-11c67ffbd6da\", \"content\": \"Create comprehensive company profile\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"1060c31b-66f7-4f69-b5aa-b7f0b8623731\", \"content\": \"Generate actionable insights and recommendations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"32f09439-effe-407d-86b0-084f86c0c990\", \"content\": \"Format final deliverable with citations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}]}], \"total_tasks\": 19, \"total_sections\": 5}, \"error\": null}}}",
      "message_id": "d08c28d0-f29e-46ec-a20d-f8d265222f8c"
    },
    {
      "role": "assistant",
      "content": "Now let me create a comprehensive workflow that will handle all your research needs. I'll set up a parameterized workflow that can research any company you specify.\n\n<function_calls>\n<invoke name=\"create_workflow\">\n<parameter name=\"name\">Comprehensive Company Research Playbook</parameter>\n<parameter name=\"description\">Complete company analysis including news, funding, leadership, c... (truncated)\n\nmessage_id \"96b69307-0f21-4009-8d03-ddd4a5b11fa0\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "96b69307-0f21-4009-8d03-ddd4a5b11fa0"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"create_workflow\", \"xml_tag_name\": \"create-workflow\", \"tool_call_id\": null, \"arguments\": {\"name\": \"Comprehensive Company Research Playbook\", \"description\": \"Complete company analysis including news, funding, leadership, competitive position, and actionable insights\", \"template\": \"# Comprehensive Company Research: {{company_name}}\\n\\n## Research Objectives\\n- Company overview and business model\\n- Recent news and developments\\n- Funding history and financial health\\n- Leadership team and key personnel\\n- Competitive landscape and market position\\n- Strategic insights and recommendations\\n\\n## Research Process\\n\\n### Phase 1: Company Fundamentals\\n1. Search for \\\"{{company_name}}\\\" company overview and business model\\n2. Research \\\"{{company_name}}\\\" recent news and announcements\\n3. Analyze \\\"{{company_name}}\\\" funding history and investment rounds\\n4. Review \\\"{{company_name}}\\\" financial performance and metrics\\n\\n### Phase 2: Leadership Analysis\\n1. Research \\\"{{company_name}}\\\" CEO and executive team\\n2. Analyze key personnel LinkedIn profiles\\n3. Research board members and advisors\\n4. Identify recent leadership changes\\n\\n### Phase 3: Market Position\\n1. Research \\\"{{company_name}}\\\" competitive landscape\\n2. Analyze market share and positioning\\n3. Identify key competitors and differentiators\\n4. Research industry trends affecting the company\\n\\n### Phase 4: Synthesis & Insights\\n1. Synthesize all findings into comprehensive profile\\n2. Generate actionable insights and recommendations\\n3. Create detailed company dossier with citations\\n4. Provide strategic recommendations\\n\\n## Deliverables\\n- Comprehensive company profile document\\n- Executive summary with key insights\\n- Competitive analysis report\\n- Strategic recommendations\", \"variables\": [{\"key\": \"company_name\", \"label\": \"Company Name\", \"required\": true}, {\"key\": \"industry_focus\", \"label\": \"Industry Focus (optional)\", \"required\": false}, {\"key\": \"research_depth\", \"label\": \"Research Depth (basic/comprehensive)\", \"required\": false}]}, \"result\": {\"success\": true, \"output\": {\"message\": \"Workflow 'Comprehensive Company Research Playbook' created successfully\", \"workflow\": {\"id\": \"d3fd2202-6256-42e3-98d3-ff4a4fc562cb\", \"name\": \"Comprehensive Company Research Playbook\", \"description\": \"Complete company analysis including news, funding, leadership, competitive position, and actionable insights\", \"is_default\": false, \"status\": \"draft\", \"steps_count\": 1, \"created_at\": \"2025-08-17T19:06:29.804107+00:00\"}}, \"error\": null}}}",
      "message_id": "28cf2a50-3774-4742-93ff-4621044f8270"
    },
    {
      "role": "assistant",
      "content": "Perfect! Now let me activate this workflow and start the research process. But first, I need to know which specific company you'd like me to research.\n\n<function_calls>\n<invoke name=\"activate_workflow\">\n<parameter name=\"workflow_id\">d3fd2202-6256-42e3-98d3-ff4a4fc562cb</parameter>\n<parameter name=\"active\">true</parameter>\n</invoke>\n</function_calls>",
      "tool_calls": null,
      "message_id": "917779ff-2ea3-4848-8845-812c4ec5816f"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"activate_workflow\", \"xml_tag_name\": \"activate-workflow\", \"tool_call_id\": null, \"arguments\": {\"workflow_id\": \"d3fd2202-6256-42e3-98d3-ff4a4fc562cb\", \"active\": true}, \"result\": {\"success\": true, \"output\": {\"message\": \"Workflow 'Comprehensive Company Research Playbook' activated successfully\", \"workflow_id\": \"d3fd2202-6256-42e3-98d3-ff4a4fc562cb\", \"status\": \"active\"}, \"error\": null}}}",
      "message_id": "876b93b3-25ed-4cad-a16a-446af90e313d"
    },
    {
      "role": "user",
      "content": "continue",
      "message_id": "79a118c1-b22c-41d8-8a05-7a64be7c7063"
    },
    {
      "role": "assistant",
      "content": "Great! The workflow is now active. Let me start by updating the first task as complete and begin the actual research process. I'll start with a comprehensive company research example to demonstrate the full workflow.\n\n<function_calls>\n<invoke name=\"update_tasks\">\n<parameter name=\"task_ids\">[\"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\"]</parameter>\n<parameter name=\"status\">completed</para... (truncated)\n\nmessage_id \"bf621916-1013-4aec-bf9a-5b93e25e0a5a\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "bf621916-1013-4aec-bf9a-5b93e25e0a5a"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"update_tasks\", \"xml_tag_name\": \"update-tasks\", \"tool_call_id\": null, \"arguments\": {\"task_ids\": [\"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\"], \"status\": \"completed\"}, \"result\": {\"success\": true, \"output\": {\"sections\": [{\"id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\", \"title\": \"Research Setup & Planning\", \"tasks\": [{\"id\": \"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\", \"content\": \"Set up company research workflow with variables\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"content\": \"Configure data sources and search parameters\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"20b5484e-fe54-4236-8f5c-e71a738a26da\", \"content\": \"Establish research methodology and sources\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}]}, {\"id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\", \"title\": \"Company Fundamentals Research\", \"tasks\": [{\"id\": \"b08bc6af-f98c-46a4-b965-8c96c963ea33\", \"content\": \"Research basic company information and overview\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"d97ef9f3-3d4a-47b0-8ae6-b8137c11ea32\", \"content\": \"Gather recent news and announcements\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"a93178e5-a6d7-4c2c-8a30-50bb78987d36\", \"content\": \"Research funding history and investment rounds\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"7ac8c5a2-8cc0-4cfa-9202-44812e7a6a33\", \"content\": \"Analyze financial performance and metrics\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}]}, {\"id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\", \"title\": \"Leadership & Personnel Analysis\", \"tasks\": [{\"id\": \"a1ae7fc4-5abd-4197-b783-319c874de228\", \"content\": \"Research CEO and executive team backgrounds\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"238940d8-d6cc-45e7-83b3-a6377137201c\", \"content\": \"Analyze key personnel LinkedIn profiles\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"ddbe449e-33a5-4675-862e-43775a0b2730\", \"content\": \"Research board members and advisors\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"643e917a-088c-4140-8426-cafb7e600faa\", \"content\": \"Identify leadership changes and transitions\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}]}, {\"id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\", \"title\": \"Market Position Analysis\", \"tasks\": [{\"id\": \"f4e90132-5b18-4b3c-9018-39c911414e9b\", \"content\": \"Research competitive landscape\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"454b1d2b-4da1-4c14-8f6e-31073a789702\", \"content\": \"Analyze market share and positioning\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"92282e56-bee1-44c2-a4ba-7679472ea5d1\", \"content\": \"Identify key competitors and differentiators\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"aa1fe7aa-c4ec-44c5-85d3-2cac70fe88c7\", \"content\": \"Research industry trends and market dynamics\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}]}, {\"id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\", \"title\": \"Insights & Profile Creation\", \"tasks\": [{\"id\": \"f78993ad-b1b9-4fc8-8896-5616c3da35c1\", \"content\": \"Synthesize all research findings\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"5e1421e0-55b6-4bdc-9d20-11c67ffbd6da\", \"content\": \"Create comprehensive company profile\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"1060c31b-66f7-4f69-b5aa-b7f0b8623731\", \"content\": \"Generate actionable insights and recommendations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"32f09439-effe-407d-86b0-084f86c0c990\", \"content\": \"Format final deliverable with citations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}]}], \"total_tasks\": 19, \"total_sections\": 5}, \"error\": null}}}",
      "message_id": "1d3c8600-fc02-4cff-900d-401d35778ffc"
    },
    {
      "role": "assistant",
      "content": "Now let me demonstrate the complete research workflow by researching **Stripe** as an example company. I'll execute each task systematically to show you how this comprehensive research process works.\n\n<function_calls>\n<invoke name=\"update_tasks\">\n<parameter name=\"task_ids\">[\"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"20b5484e-fe54-4236-8f5c-e71a738a26da\"]</parameter>\n<parameter name=\"... (truncated)\n\nmessage_id \"7c9dc8a4-641f-4db1-92d4-8fd39752031e\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "7c9dc8a4-641f-4db1-92d4-8fd39752031e"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"update_tasks\", \"xml_tag_name\": \"update-tasks\", \"tool_call_id\": null, \"arguments\": {\"task_ids\": [\"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"20b5484e-fe54-4236-8f5c-e71a738a26da\"], \"status\": \"completed\"}, \"result\": {\"success\": true, \"output\": {\"sections\": [{\"id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\", \"title\": \"Research Setup & Planning\", \"tasks\": [{\"id\": \"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\", \"content\": \"Set up company research workflow with variables\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"content\": \"Configure data sources and search parameters\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"20b5484e-fe54-4236-8f5c-e71a738a26da\", \"content\": \"Establish research methodology and sources\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}]}, {\"id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\", \"title\": \"Company Fundamentals Research\", \"tasks\": [{\"id\": \"b08bc6af-f98c-46a4-b965-8c96c963ea33\", \"content\": \"Research basic company information and overview\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"d97ef9f3-3d4a-47b0-8ae6-b8137c11ea32\", \"content\": \"Gather recent news and announcements\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"a93178e5-a6d7-4c2c-8a30-50bb78987d36\", \"content\": \"Research funding history and investment rounds\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"7ac8c5a2-8cc0-4cfa-9202-44812e7a6a33\", \"content\": \"Analyze financial performance and metrics\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}]}, {\"id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\", \"title\": \"Leadership & Personnel Analysis\", \"tasks\": [{\"id\": \"a1ae7fc4-5abd-4197-b783-319c874de228\", \"content\": \"Research CEO and executive team backgrounds\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"238940d8-d6cc-45e7-83b3-a6377137201c\", \"content\": \"Analyze key personnel LinkedIn profiles\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"ddbe449e-33a5-4675-862e-43775a0b2730\", \"content\": \"Research board members and advisors\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"643e917a-088c-4140-8426-cafb7e600faa\", \"content\": \"Identify leadership changes and transitions\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}]}, {\"id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\", \"title\": \"Market Position Analysis\", \"tasks\": [{\"id\": \"f4e90132-5b18-4b3c-9018-39c911414e9b\", \"content\": \"Research competitive landscape\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"454b1d2b-4da1-4c14-8f6e-31073a789702\", \"content\": \"Analyze market share and positioning\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"92282e56-bee1-44c2-a4ba-7679472ea5d1\", \"content\": \"Identify key competitors and differentiators\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"aa1fe7aa-c4ec-44c5-85d3-2cac70fe88c7\", \"content\": \"Research industry trends and market dynamics\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}]}, {\"id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\", \"title\": \"Insights & Profile Creation\", \"tasks\": [{\"id\": \"f78993ad-b1b9-4fc8-8896-5616c3da35c1\", \"content\": \"Synthesize all research findings\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"5e1421e0-55b6-4bdc-9d20-11c67ffbd6da\", \"content\": \"Create comprehensive company profile\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"1060c31b-66f7-4f69-b5aa-b7f0b8623731\", \"content\": \"Generate actionable insights and recommendations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"32f09439-effe-407d-86b0-084f86c0c990\", \"content\": \"Format final deliverable with citations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}]}], \"total_tasks\": 19, \"total_sections\": 5}, \"error\": null}}}",
      "message_id": "27f25d61-6912-4755-a09a-e615738aa325"
    },
    {
      "role": "assistant",
      "content": "Now let me begin the actual research on Stripe. I'll start with basic company information and work through each section systematically.\n\n<function_calls>\n<invoke name=\"web_search\">\n<parameter name=\"query\">Stripe company overview business model revenue 2024 2025</parameter>\n<parameter name=\"num_results\">15</parameter>\n</invoke>\n</function_calls>",
      "tool_calls": null,
      "message_id": "4ab32962-1f27-4c6c-b680-bbd0c6d4ebb9"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"web_search\", \"xml_tag_name\": \"web-search\", \"tool_call_id\": null, \"arguments\": {\"query\": \"Stripe company overview business model revenue 2024 2025\", \"num_results\": 15}, \"result\": {\"success\": true, \"output\": {\"query\": \"Stripe company overview business model revenue 2024 2025\", \"follow_up_questions\": null, \"answer\": \"Stripe is a leading financial infrastructure company that processes payments for millions of businesses worldwide, handling over $1.4 trillion in total payment volume in 2024, representing approximately 1.3% of global GDP. The company operates on a commission-based business model, typically charging 2.9% plus 30 cents per transaction for domestic cards, with custom pricing available for enterprise clients. Stripe achieved profitability in 2024 and generated an estimated $18-19 billion in revenue, with projections suggesting continued growth into 2025. The platform serves over 1.3 million active websites globally across 46+ countries, supporting 135+ currencies and payment methods, with more than 100 enterprise customers each processing over $1 billion annually through Stripe. Key revenue drivers include their core payment processing services, subscription billing tools (which reached a $500 million revenue run rate and serve over 300,000 companies), and emerging products like Stripe Connect, Stripe Issuing, and AI-powered automation tools. The company is currently valued at approximately $70-91.5 billion and has recently made strategic acquisitions including Bridge, a stablecoin platform purchased for $1.1 billion in 2025, signaling expansion into cryptocurrency infrastructure and positioning for future growth in digital financial services.\", \"images\": [\"https://helplama.com/wp-content/uploads/2023/06/stripe-growth.jpg\", \"https://businessmodelanalyst.com/wp-content/uploads/2021/07/Stripe-Business-Model-Canvas.jpeg\", \"https://lumosbusiness.com/wp-content/uploads/2022/10/Stripe_Business_Model_Canvas.jpg\", \"https://images.ctfassets.net/kepwgvmwgwnt/sWy772Vs4SZRWREoDAnhF/8b5c6b555540daa1db3367aba2013700/stripe-mrr-dashboard.png\", \"https://kinsta.com/wp-content/uploads/2022/09/Stripe-Revenue-per-Year.png\"], \"results\": [{\"url\": \"https://coinlaw.io/stripe-statistics/\", \"title\": \"Stripe Statistics 2025: Usage, Revenue, and Market Share\", \"content\": \"Stripe generated $19.4 billion in revenue in 2025, reflecting a 17% year-over-year growth from 2024.\\n   Subscription tools contributed $3.4 billion in revenue, solidifying their role as a core business driver.\\n   The company posted a 10.6% net profit margin in 2025, maintaining profitability amid fintech competition.\\n   Annual recurring revenue (ARR) exceeded $6.1 billion in 2025, fueled by enterprise-scale adoption. [...] Stripe\\u2019s AI automation tools reduced manual workflows by 26% in 2025, improving operational efficiency for businesses.\\n   The platform\\u2019s custom checkout APIs saw 41% growth in enterprise adoption, enabling highly personalized user experiences.\\n   Stripe Issuing grew by 58%, processing over $13.4 billion in virtual and physical card transactions in 2025.\\n   Stripe Connect generated $340 million in new revenue streams in 2025, driven by platform economy expansion. [...] In 2025, Stripe processed $1.05 trillion in total payment volume, marking a 16% year-over-year increase.\\n   The platform handled $58 billion in cross-border payments, supported by smarter FX and localized payout tools.\\n   Apple Pay and Google Pay transactions grew 31%, now making up 38% of Stripe\\u2019s mobile payment volume.\\n   Stripe\\u2019s fraud system blocked $2.3 billion in fraudulent activity in 2025, showcasing AI-powered risk prevention.\", \"score\": 0.94761914, \"raw_content\": null}, {\"url\": \"https://assets.stripeassets.com/fzn2n1nzq965/2pt3yIHthraqR1KwXgr98U/df10795aac0205789956c89e0dfc4f1a/Stripe-annual-letter-2024.pdf\", \"title\": \"[PDF] Stripe-annual-letter-2024.pdf\", \"content\": \"February 27, 2025 Dear Stripe community, Businesses on Stripe generated $1.4 trillion in total payment volume in 2024, up 38% from the prior year, and reaching a scale equivalent to around 1.3% of global GDP. We attribute this year\\u2019s rapid growth in part to our long-standing investments in building machine learning and artificial intelligence into our products. These bets continue to pay off, increasing revenue for existing customers, encouraging more businesses to switch to Stripe, and helping [...] new companies reach significant scale unprecedentedly quickly\\u2014more on all of those points later. Future growth rates will fluctuate (2024 was an unusually good year), but we\\u2019re as enthusiastic as ever about the long-term trends in the internet economy. Stripe was profitable in 2024, and we expect to be so in 2025 and beyond. Durable profitability allows us to plow back much of our operating earnings into research and development. In each of the last six years, Stripe has reinvested a much [...] right by both and , recognizing its power and maturity. As we forecast in our , Stripe\\u2019s Revenue and Finance Automation Suite, with Billing at its core, has now passed a $500 million revenue run rate. Billing is being used by more than 300,000 companies, manages nearly 200 million active subscriptions, and is emerging as the revenue engine of the AI era. The businesses on Stripe span every chromosome of the economic genome, from top corporate leaders (half of the Fortune 100 uses Stripe) to\", \"score\": 0.9377661, \"raw_content\": null}, {\"url\": \"https://capitaloneshopping.com/research/stripe-statistics/\", \"title\": \"Stripe Statistics (2025): Revenue, Market Share & Growth ...\", \"content\": \"All Reports\\n\\nHighlights. 1.31 million active websites globally use Stripe, with an average of 1,000 companies joining Stripe each day.\\n\\n In 2024, Stripe processed a global payment volume of over $1.4 trillion.\\n As of May 2025, Stripe is valued at $91.5 billion.\\n In 2024, Stripe had an estimated net revenue of $5.1 billion.\\n Stripe holds a 20.8% market share in payment processing software.\\n Stripe has raised a lifetime total of $9.4 billion in investments.\\n\\n## Stripe User Statistics [...] PayPal services are used by 11.9 million websites worldwide; PayPal services are used by 805% more websites than Stripe.\\n In 2024, Stripe competitors Square and Adyen were valued at $40.7 billion and $61.9 billion respectively, compared to Stripe\\u2019s $91.5 billion valuation.\\n In 2024, Stripe recorded a net revenue of $5.1 billion 136% more than Adyen\\u2019s net revenue of $2.16 billion.\\n Square has raised $9.4 billion in investments and Adyen only $266 million.\\n\\n## Stripe Black Friday Statistics [...] | Ranking | Company | 2024 Revenue  (in billions) |\\n| 1 | Walmart | $681.0 |\\n| 2 | Amazon | $650.3 |\\n| 3 | Apple | $400.4 |\\n| 4 | Samsung | $219.6 |\\n| 5 | Walgreens | $151.9 |\\n| 6 | AliExpress | $130.4 |\\n| 7 | Target | $106.6 |\\n| 8 | Chewy | $11.86 |\\n| 9 | Wayfair | $11.85 |\\n| 10 | ASOS | $3.81 |\\n\\n## Stripe User Growth Statistics\\n\\nStripe is one of the fastest-growing online payment processors.\", \"score\": 0.903098, \"raw_content\": null}, {\"url\": \"https://coinlaw.io/paypal-vs-stripe-statistics/\", \"title\": \"PayPal vs. Stripe Statistics 2025: User Base, Fees & Global ...\", \"content\": \"PayPal\\u2019s total revenue reached $32.1 billion in 2025, growing 8.6% YoY.\\n Stripe\\u2019s projected revenue for 2025 stands at $18.9 billion, up 11.2% from $17 billion in 2024.\\n Stripe posted its first-ever quarterly profit in Q1 2025, driven by enterprise growth and infrastructure cost optimizations.\\n PayPal\\u2019s operating margin held steady at 19.4%, while Stripe grew to 12.2% in 2025.\\n Stripe Treasury became profitable for the first time, handling $3.1 billion in business balances and payouts. [...] PayPal\\u2019s revenue from Braintree grew 13.5%, contributing $7.4 billion to its topline.\\n Stripe\\u2019s Revenue Intelligence tools brought in $1.2 billion in new upsell-driven income in 2025.\\n PayPal saw net income of $5.3 billion in 2025, while Stripe posted an estimated net income of $2.6 billion.\\n PayPal\\u2018s recurring revenue streams made up 42% of total revenue, indicating successful diversification efforts. [...] PayPal\\u2018s peer-to-peer (P2P) volume dropped slightly to $315 billion in 2025, as Venmo\\u2019s growth plateaued.\\n Stripe\\u2019s recurring billing volume reached $190 billion in 2025, fueled by its Billing and Invoicing APIs.\\n The average transaction value on PayPal was $84.60 per payment in 2025.\\n The average Stripe transaction came in higher at $98.10, due to its B2B-heavy customer base.\\n PayPal processed over 5.4 billion transactions in Q1\\u2013Q2 of 2025 alone.\", \"score\": 0.8967681, \"raw_content\": null}, {\"url\": \"https://www.chargeflow.io/blog/stripe-statistics\", \"title\": \"Stripe Stats for 2025: The Most Accurate Stripe Statistics ...\", \"content\": \"1.  Stripe generated $1.4 trillion in total payment volume in 2024, up 38% from the prior year.\\n2.  Merchants in 50 countries use Stripe to accept and process payments.\\n3.  More than 100 Stripe customers process $1 billion or more yearly (40 in 2020).\\n4.  4,776,230 million websites have used Stripe to accept payment.\\n5.  Stripe is currently valued at $70 billion, making it one of the most valuable privately-held companies on the planet. [...] ### Stripe Black Friday Cyber Monday Statistics\\n\\nStripe\\u2019s BFCM stats are always a thing for the books and 2024 was no different. Businesses processed more than $31 billion on Stripe from Black Friday through Cyber Monday, the company proclaimed. That\\u2019s a notable uptick from $18.6 billion in 2023. [...] Image 31\\n\\nStripe Revenue Statistics\\n-------------------------\\n\\nStripe's business model is charging a commission for every transaction processed, typically at a standard rate of 2.9% + 30 cents (for domestic cards) of the transaction. However, they offer custom pricing options for enterprise clients. Stripe Radar, a fraud protection tool, charges 5 to 7 cents for each screened transaction.\", \"score\": 0.8452414, \"raw_content\": null}, {\"url\": \"https://sacra.com/c/stripe/valuation/\", \"title\": \"Stripe: $91.50B valuation [2025 - Sacra\", \"content\": \"| Bull \\ud83d\\ude80 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $23B | $30B | $40B | $53B | $71B |\\n| Growth | 25.00% | 27.69% | 30.00% | 32.31% | 33.81% | 35% |\\n\\nOur bear, base, and bull cases for Stripe reflect varying scenarios of competitive pressure, market penetration, and ecosystem expansion success [...] | Bear \\ud83d\\udc3b | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $22B | $26B | $31B | $36B | $42B |\\n| Growth | 25.00% | 22.31% | 20.00% | 17.69% | 16.19% | 15% |\\n\\n| Base \\ud83d\\udcc8 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $22B | $27B | $33B | $40B | $48B |\\n| Growth | 25.00% | 23.66% | 22.50% | 21.34% | 20.60% | 20% | [...] Current revenue of $18B and 25% growth rate demonstrate Stripe's continued market strength, though growth has moderated from previous years' rapid expansion. This positions them well among B2B payment processors while reflecting a maturing business in the competitive fintech landscape.\\n\\n| Multiple | Valuation |\\n| --- | --- |\\n| 1x | $18B |\\n| 5x | $90B |\\n| 10x | $180B |\\n| 15x | $270B |\\n| 25x | $450B |\", \"score\": 0.80698997, \"raw_content\": null}, {\"url\": \"https://samboboev.medium.com/deep-dive-stripe-vs-adyen-comparing-2024-performance-ab585de2d27d\", \"title\": \"Deep Dive: Stripe vs. Adyen: Comparing 2024 Performance\", \"content\": \"In terms of revenue, Adyen\\u2019s business model focuses on \\u201cNet Revenue,\\u201d which excludes pass-through costs. Adyen\\u2019s net revenue for 2024 was \\u20ac1,996.1 million, a 23% YoY increase\\u200b. Stripe, as a private company, does not publicly break out revenue figures in its community letter; however, the company confirmed that it was profitable in 2024\\u200b. Stripe\\u2019s profitability in 2024 marked a notable milestone, enabling it to \\u201cplow back\\u201d operating earnings into R&D\\u200b. Adyen also demonstrated strong [...] Both Stripe and Adyen reported strong growth in 2024, with significant transaction volumes and improving profitability. Total Payment Volume (TPV) on Stripe\\u2019s platform reached $1.4 trillion for 2024, a 38% year-over-year (YoY) increase\\u200b. This makes Stripe responsible for about 1.3% of global GDP in payments processing\\u200b. Adyen\\u2019s processed volume was of a similar magnitude at \\u20ac1.29 trillion in 2024, which was up 33% YoY\\u200b. Adyen\\u2019s management notes that excluding a single large client (eBay), [...] This week, we\\u2019re diving into the battle of fintech giants: Stripe vs. Adyen. Both payment powerhouses had a stellar 2024, but they took different paths to success. Stripe saw $1.4 trillion in Total Payment Volume (TPV), growing 38% YoY, while Adyen wasn\\u2019t far behind with \\u20ac1.29 trillion processed (+33% YoY). Adyen maintained its 50% EBITDA margin, while Stripe finally hit full-year profitability, proving its business model can scale.\", \"score\": 0.77994305, \"raw_content\": null}, {\"url\": \"https://www.demandsage.com/stripe-statistics/\", \"title\": \"Stripe Revenue Statistics (2025): Market Share Data\", \"content\": \"Initially launched in the United States, Stripe has now expanded across 46 countries. However, the valuation of Stripe was reduced from $95 billion to $70 billion in the span of 3 years.\\n\\nInterested to know more details about Stripe for 2025?\\n\\nI have compiled the latest Stripe Statistics of 2025 in this article. So, let\\u2019s get into the details!\\n\\nStripe Statistics 2025: At A Glance\\n--------------------------------------- [...] According to John Collison (co-founder of Stripe), the company has over 40 businesses with a $1 billion payment volume each.\\n\\nFurther, according to the company data, there is two times increase in the payment volume of the Stripe enterprise clients.\\n\\nSource:Backlinko, Twitter, Stripe.\\n\\nHow Does Stripe Make Money?\\n------------------------------- [...] The stripe model charges a commission on every successful transaction. The standard rate of the charges made by Stripe is 2.9% of the transaction + 25 cents.\\n\\nHowever, the transaction fees or the commission rate may vary according to the plan the client has opted for.\\n\\nAdditionally, enterprise clients get a custom pricing option that tailors the features according to their requirements.\", \"score\": 0.76194656, \"raw_content\": null}, {\"url\": \"https://stripe.com/en-li/annual-updates/2024?__hsfp=1096049464&__hssc=106715356.1.1668124800362&__previewId&__s=1ngpkifa1w8cdaucpyke\", \"title\": \"Stripe's 2024 annual letter\", \"content\": \"Start-up incorporation\\n\\nCarbon removal\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Stripe\\u2019s 2024 annual letter\\n\\n2024 was a good year for the internet economy and for our users: in aggregate, their revenue processed on Stripe grew seven times faster than the revenue of companies in the S&P 500. [...] # Stripe logo\\n\\n# Global payments\\n\\nOnline payments\\n\\nFraud prevention\\n\\nAcceptance optimisations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation\\n\\nSales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n# More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStart-up incorporation\\n\\n# By stage\\n\\n# By business model [...] # By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\n# Get started\\n\\n# Guides\\n\\n# Global payments\\n\\nOnline payments\\n\\nFraud prevention\\n\\nAcceptance optimisations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation\\n\\nSales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\", \"score\": 0.7578844, \"raw_content\": null}, {\"url\": \"https://backlinko.com/stripe-users\", \"title\": \"Stripe Revenue and Growth Statistics (2024) - Backlinko\", \"content\": \"That\\u2019s all for my collection of Stripe stats for 2024. Stripe has quickly grown into one of the major players in the fintech industry, processing one trillion dollars annually.\\n\\nHopefully, you found these facts and figures interesting. As a private company, Stripe doesn\\u2019t disclose a lot of juicy data. But I did my best to collect what was out there.\\n\\nNext-level SEO training and link building strategies\\n\\n#### Learn\\n\\n#### Company\\n\\n#### Connect\\n\\n\\u00a9 2025 Backlinko is a Trademark of Semrush Inc [...] Here\\u2019s the breakdown of the number of available payment methods supported by Stripe in each category:\\n\\n| Category | Number of payment methods |\\n| --- | --- |\\n| Cards | 10 |\\n| Bank debits | 6 |\\n| Bank redirects | 10 |\\n| Bank transfers | 5 |\\n| Buy now, pay later | 5 |\\n| Real-time payments | 4 |\\n| Vouchers | 4 |\\n| Wallets | 15 |\\n\\nSource: Stripe\\n\\n## How Does Stripe Make Money?\\n\\nThe Stripe business model is built on charging a commission per successful transaction. [...] ## Stripe Enterprise Growth\\n\\nStripe has over 100 companies that process more than $1 billion via Stripe on an annual basis each.\\n\\nNumber of large customers processing more than $1 billion each has more than doubled from more than 40 firms reported earlier in 2020.\\n\\nSources: Stripe, X/Twitter\\n\\n## Who Founded Stripe?\\n\\nStripe was founded in 2010 by brothers Patrick and John Collison.\", \"score\": 0.7402687, \"raw_content\": null}, {\"url\": \"https://helplama.com/stripe-revenue-users-statistics/\", \"title\": \"Stripe Revenue and Users Statistics 2025\", \"content\": \"Stripe Revenue and Users Statistics 2025\\n========================================\\n\\nImage 2: Stripe Revenue Users Stats\\n\\nLast Updated:August 2025\\n\\nStripe, the leading online payments platform, has been on an incredible trajectory, revolutionizing the way businesses transact in the digital age. As we dive into the year 2025, it\\u2019s time to uncover the astonishing Stripe statistics that highlight its unstoppable rise in revenue and user engagement. [...] Stripe Revenue 2024\\n-----------------------\\n\\n### Stripe Annual Revenue\\n\\nStripe revenue increased by 20% to $14 billion in 2022 from $12 billion in 2021.\\n\\nSpeaking of Stripe\\u2019s revenue growth, the payment giant\\u2019s gross revenue has seen a compound annual growth rate (CAGR) of 69.52% since 2017.\\n\\nFinancial yearRevenue\\n2014$40 million\\n2016$450 million\\n2017$1 billion\\n2018$1.5 billion\\n2019$2 billion\\n2020$7.4 billion\\n2021$12 billion\\n2022$14 billion\\n\\nSource: Backlinko, Sacra.com\\n\\nImage 3: Stripe Growth [...] When current customers are combined with historical ones, the number is estimated at 3,820,274.\\n\\nSource: Builtwith.com\\n\\n### What Companies Use Stripe?\\n\\nOver 100 companies use Stripe to process over $1 billion annually.\\n\\nCompany NameWebsiteRevenueLocation\\nFlexport www.flexport.com$1 Billion United States\\nInstacart www.instacart.com$1 Billion United States\\nShopify www.shopify.com$4 Billion Canada\\nLyft www.lyft.com$2 Billion United States\\nWayfair www.wayfair.com$14 Billion United States\", \"score\": 0.72465646, \"raw_content\": null}, {\"url\": \"https://www.forbes.com/companies/stripe/\", \"title\": \"Stripe | Company Overview & News - Forbes\", \"content\": \"Business Services & Supplies\\n\\n#183\\n\\nAmerica's Top Private Companies (2024)\\n\\nBusiness Services & Supplies\\n\\n#3\\n\\nCloud 100 (2024)\\n\\nFinancial Infrastructure\\n\\nMore Lists\\n\\n### Stripe Financial Summary\\n\\nSelect list year\\n\\n2024\\n\\n2023\\n\\n2022\\n\\nRevenue\\n\\n$3.5B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRevenue\\n\\n$3.1B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRevenue\\n\\n$2.5B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRelated People & Companies\\n\\nImage 1: John Collison\\n\\nJohn Collison\\n\\nEmployed by Company\\n\\nView Profile\\n\\nImage 2: Peregrine\\n\\nPeregrine [...] Founded in 2009 by Irish brothers Patrick and John Collison, Stripe processes payments for online businesses. In 2024, its subscription-billing product reached more than 300,000 business customers, and it agreed to acquire Bridge, which makes infrastructure software for cryptocurrency-based stable coins (tokens that are typically pegged to the value of the U.S. dollar), for $1.1 billion. Stripe counts Nvidia, OpenAI and Pepsi among its customers. [...] Stripe | Company Overview & News\\n\\n===============\\n\\n, for $1.1 billion. Stripe counts Nvidia, OpenAI and Pepsi among its customers. \\nFunding: $2.2 billion from Sequoia, General Catalyst, Founders Fund, Thrive and others.\\n\\nLatest valuation: $70 billion, based on a September 2024 sha... Read More\", \"score\": 0.67192334, \"raw_content\": null}, {\"url\": \"https://en.wikipedia.org/wiki/Stripe,_Inc.\", \"title\": \"Stripe, Inc. - Wikipedia\", \"content\": \"On 4 February 2025, Stripe announced it had acquired Bridge.( The acquisition of the two-year-old stablecoin platform company is valued at $1.1 billion.(\\n\\nInvestments\\n-----------\\n\\n[edit] [...] On 26 July 2024, Stripe acquired a Merchant of Record services start-up Lemon Squeezy&action=edit&redlink=1 \\\"Lemon Squeezy (startup) (page does not exist)\\\").(\\n\\nSee also\\n--------\\n\\n[edit]\\n\\n   Electronic commerce\\n   List of online payment service providers\\n   Payment gateway\\n   Payment service provider\\n   Subscription business model\\n\\nReferences\\n----------\\n\\n[edit] [...] 80.   ^\\\"Stripe completes Bridge acquisition\\\". _Stripe, Inc_. 4 February 2025. Retrieved 8 April 2025.\\n81.   ^Sigalos, MacKenzie (4 February 2025). \\\"Stripe closes $1.1 billion Bridge deal, prepares for aggressive stablecoin push\\\". _CNBC_. Retrieved 8 April 2025. Stripe closed its $1.1 billion purchase of Bridge Network.\", \"score\": 0.5923135, \"raw_content\": null}, {\"url\": \"https://stripe.com/\", \"title\": \"Stripe | Financial Infrastructure to Grow Your Revenue\", \"content\": \"- Jobs\\n- Newsroom\\n\\n- Stripe Press\\n- Become a partner\\n\\nSessions 2025\\n\\u00b7\\nWatch on demand\\n\\n# Financial infrastructure to grow your revenue\\n\\nFinancial infrastructure to grow your revenue\\n\\nFinancial infrastructure to grow your revenue\\n\\nJoin the millions of companies that use Stripe to accept payments online and in person, embed financial services, power custom revenue models, and build a more profitable business.\\n\\nStart now\\nContact sales\\n\\nStart now\\nContact sales\\n\\nRocket Rides\\n\\nSearch\\n\\nToday [...] # The backbone for global commerce\\n\\nStripe makes moving money as easy and programmable as moving data.\\nOur teams are based in offices around the world and we process hundreds of billions of dollars each year for ambitious businesses of all sizes.\\n\\n# 500M+\\n\\nAPI requests per day, peaking at 13,000 requests a second.\\n\\n# 99.999%\\n\\nhistorical uptime for Stripe services.\\n\\n# 90%\\n\\nof U.S. adults have bought from businesses using Stripe.\\n\\n# 135+\\n\\ncurrencies and payment methods supported.\", \"score\": 0.5360463, \"raw_content\": null}, {\"url\": \"https://onstrategy.eu/stripes-2025-letter-explained/\", \"title\": \"Stripe's 2025 Letter explained - onStrategy\", \"content\": \"Stripe\\u2019s acquisition of Bridge, the leading stablecoin orchestration platform, signals a massive shift in financial infrastructure. Stablecoins make money movement cheaper, faster, and programmable, and Stripe is betting that they will become a core part of business transactions worldwide. While stablecoins have been largely a crypto-native phenomenon, Stripe\\u2019s push into the space suggests that they\\u2019re about to become mainstream for corporate treasury, global payroll, and real-world commerce, [...] , Traxero (for tow truck operators), and Clio (for law firms), all examples of vertical SaaS businesses that act as financial platforms for their industries. This is a fundamental shift: Stripe it\\u2019s enabling software-driven economies at every level, including small businesses that previously couldn\\u2019t access this kind of technology.\\n\\n7/ Stablecoins are Stripe\\u2019s next big bet, because money itself is changing.\", \"score\": 0.5106191, \"raw_content\": null}], \"response_time\": 9.53, \"request_id\": \"adf9499f-4b2e-4a4d-8ca5-1ae55f579f20\"}, \"error\": null}}}",
      "message_id": "e5e48ad5-fa67-4982-83f7-1bcbe9aabe5e"
    },
    {
      "role": "assistant",
      "content": "<function_calls>\n<invoke name=\"web_search\">\n<parameter name=\"query\">Stripe recent news announcements 2024 2025 acquisitions partnerships</parameter>\n<parameter name=\"num_results\">10</parameter>\n</invoke>\n</function_calls>",
      "tool_calls": null,
      "message_id": "c6a9b015-c903-4ad3-b768-17b665130f6b"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"web_search\", \"xml_tag_name\": \"web-search\", \"tool_call_id\": null, \"arguments\": {\"query\": \"Stripe recent news announcements 2024 2025 acquisitions partnerships\", \"num_results\": 10}, \"result\": {\"success\": true, \"output\": {\"query\": \"Stripe recent news announcements 2024 2025 acquisitions partnerships\", \"follow_up_questions\": null, \"answer\": \"Stripe has been highly active with acquisitions and partnerships throughout 2024 and 2025, completing several major deals that expand its capabilities across different payment sectors. The company acquired Bridge, a stablecoin platform, in February 2025 in what represents Stripe's largest acquisition to date, positioning the company to leverage the $15.6 trillion stablecoin market that moved in 2024. Other significant acquisitions include Orum, a payments orchestration platform for instant payment capabilities, acquired in July 2025, and Lemon Squeezy, which was acquired in 2024 to enhance digital product sales capabilities. Stripe also announced plans to acquire Recko to automate payments reconciliation and OpenChannel to expand its platform offerings. On the partnership front, Stripe has deepened its collaboration with NVIDIA to enhance AI-powered capabilities and expand global access to NVIDIA's AI platform, while also establishing partnerships with Klarna to help online businesses grow revenue, Salesforce for global commerce, and Shopify to enable stablecoin payments for millions of merchants. The company processed over $1.4 trillion in total payment volume in 2024, representing 38% growth, and now serves half of the Fortune 100 companies, demonstrating its evolution from a startup-focused platform to a major enterprise player.\", \"images\": [\"https://images.stripeassets.com/fzn2n1nzq965/1ZK0vLagjDOXEdwyApsNIQ/22fd18c98bcb8ca7d550ebf6c08678d4/Newsroom_carousel-bridge_1080x736.jpg?q=80&w=1080&fm=jpg\", \"https://7834608.fs1.hubspotusercontent-na1.net/hub/7834608/hubfs/Slide4-May-13-2024-11-41-48-2955-AM.png?width=1200&length=1200&name=Slide4-May-13-2024-11-41-48-2955-AM.png\", \"https://images.stripeassets.com/fzn2n1nzq965/kxTRUol9vTSEnrwOQAbpE/98effe9be044e01843344712996333f4/Clear___Stripe_4000x2000_v2.png?q=80&w=552&fm=jpg\", \"https://media.licdn.com/dms/image/v2/D4D10AQHe4_8XoctNJw/image-shrink_1280/image-shrink_1280/0/1732536005800?e=2147483647&v=beta&t=W7jPeXOVfHtljCWml648GKWc2y8_5loZjNui1pbsocg\", \"https://images.stripeassets.com/fzn2n1nzq965/77hV1QeIA0vzLaAzhgtPtQ/458eca960412cd1b4fd7d519a0f98f12/NRP_NVIDIA.png?q=80&w=1080&fm=jpg\"], \"results\": [{\"url\": \"https://www.fintechfutures.com/m-a/stripe-acquires-orum\", \"title\": \"Stripe acquires paytech start-up Orum - FinTech Futures\", \"content\": \"Orum is a payments orchestration platform that connects businesses to instant payment capabilities via API.\\n\\nPicture of Tyler Pathe\\n\\nJuly 24, 2025\\n\\nOrum - FinTech News\\n\\nOrum, a payments orchestration platform connecting businesses to instant payment capabilities via API, is being acquired by US fintech giant Stripe for an undisclosed sum. [...] .png?width=300&auto=webp&quality=80&disable=upscale)\\nFinTech Futures\\n\\n##### Events\\n\\n##### Resources\\n\\n##### Working With Us\\n\\n##### Join Us\\n\\n##### Follow Us\\n\\nInforma\\n\\nCopyright \\u00a9 2025 Informa PLC. Informa PLC is registered in England and Wales with company number 8860726 whose registered and head office is 5 Howick Place, London, SW1P 1WG. [...] In a blog post announcing the acquisition, Kirkpatrick states: \\\"After thoughtful consideration, it became clear that we have a rare opportunity to accelerate Orum\\u2019s mission and greatly increase our impact by becoming part of Stripe, which processed more than $1.4 trillion in total payment volume in 2024.\\\"\", \"score\": 0.776708, \"raw_content\": null}, {\"url\": \"https://www.chargeflow.io/blog/stripe-statistics\", \"title\": \"Stripe Stats for 2025: The Most Accurate Stripe Statistics Online\", \"content\": \"### Full List of Stripe Acquisitions to Date\\n\\nStripe has acquired 20 SaaS, Enterprise Software, and Enterprise Tech companies in the \\u00c2 U.S. and other countries. In classic Stripe fashion, most of the investment amounts are undisclosed. Here are 19 Stripe acquisitions that have currently reached our radar:\\n\\n### 20 Hottest Stripe Investments to Date [...] Stripe\\u00e2\\u0080\\u0099s BFCM stats are always a thing for the books and 2024 was no different. Businesses processed more than $31 billion on Stripe from Black Friday through Cyber Monday, the company proclaimed. That\\u00e2\\u0080\\u0099s a notable uptick from $18.6 billion in 2023.\\n\\n## Stripe Investment & Acquisition Stats\\n\\nStripe has carried out 18 acquisitions since 2013. The payment giant has also funded several tech companies since 2017. Some of these strategic investments reach hundreds of millions of U.S. dollars. [...] Stripe continues to fire on all cylinders after achieving unicorn status in just three years. The payment giant is one of the fastest-growing fintechs worldwide, with customers including Apple, Alaska Airlines, Best Buy, Lotus Cars, Microsoft, Uber, and Zara.\\n\\n## Stripe Revenue Statistics\", \"score\": 0.7371019, \"raw_content\": null}, {\"url\": \"https://a16z.com/newsletter/what-stripes-acquisition-of-bridge-means-for-fintech-and-stablecoins-april-2025-fintech-newsletter/\", \"title\": \"What Stripe's Acquisition of Bridge Means for Fintech and Stablecoins\", \"content\": \"###### TOP SUGGESTIONS\\n\\n# What Stripe\\u2019s Acquisition of Bridge Means for Fintech and Stablecoins (April 2025 Fintech Newsletter)\\n\\nJames da Costa and Sam Broner\\n\\n#### What Stripe\\u2019s Acquisition of Bridge Means for Fintech and Stablecoins [...] their merchants a cheaper alternative to credit card payments. In 2024, Stripe grew 38% \\u2014 surpassing $1.4 trillion in total paymentsvolume; the Bridge acquisition could therefore fuel further global growth through stablecoins. [...] In February, Stripe completed its acquisition of stablecoin platform Bridge \\u2014 its largest acquisition to date \\u2014 following the reintroduction of crypto payments on Stripe for U.S. businesses last year. More broadly, after years of steady stablecoin growth, this acquisition represents the payment industry\\u2019s first big acknowledgement that stablecoins can go further mainstream. In 2024, stablecoins moved $15.6 trillion in value \\u2014 putting stablecoin transaction volume on par with Visa\\u2019s.\", \"score\": 0.72395444, \"raw_content\": null}, {\"url\": \"https://stripe.com/newsroom/news\", \"title\": \"Stripe Newsroom: The Latest News & Announcements\", \"content\": \"# Stripe Capital drives sharp increases in small businesses\\u2019 revenue growth\\n\\n## Corporate\\n\\n# Stripe acquires OpenChannel\\n\\n## Product\\n\\n# Stripe expands Stripe Terminal across Europe\\n\\n## Corporate\\n\\n# Matt Huang joins Stripe\\u2019s board of directors\\n\\n## Product\\n\\n# Klarna partners with Stripe to help online businesses grow their revenue\\n\\n## Corporate\\n\\n# Stripe will acquire Recko to help internet businesses automate payments reconciliation\\n\\n## Product [...] # Stripe announces expanded interoperability, major upgrades to Stripe Billing, new partnerships with NVIDIA and Pepsi, and more\\n\\n## Corporate\\n\\n# Stripe deepens collaboration with NVIDIA to enhance Stripe\\u2019s AI-powered capabilities and expand global access to NVIDIA\\u2019s AI platform\\n\\n## Product\\n\\n# Stripe expands in Toronto to accelerate Canadian business growth\\n\\n## Product\\n\\n# Enercare chooses Stripe as its strategic payments partner\\n\\n## Product [...] ## Corporate\\n\\n# Stripe launches world\\u2019s first large-scale carbon removal purchase tool\\n\\n## Corporate\\n\\n# Stripe will acquire Paystack to accelerate online commerce across Africa\\n\\n## Corporate\\n\\n# Stripe named a Leader by Forrester, strikes global commerce partnership with Salesforce\\n\\n## Corporate\\n\\n# Dhivya Suryadevara to join Stripe as Chief Financial Officer\\n\\n## Corporate\\n\\n# Stripe hires Mike Clayville as Chief Revenue Officer\\n\\n## Product\", \"score\": 0.7140107, \"raw_content\": null}, {\"url\": \"https://stripe.com/blog/top-product-updates-sessions-2025\", \"title\": \"Our top product updates from Sessions 2025\", \"content\": \"Stripe Profiles\\n\\nIn addition to sharing updates from our four core product areas, we also announced several other features, products, and programs that help you grow. We announced a new API that lets you embed agentic commerce directly into your applications, unveiled Stripe Verified, a new credential for an accredited Stripe business, and previewed a new networking capability that lets you easily transact with other businesses on Stripe. Here\\u2019s what\\u2019s new: [...] More than 300,000 businesses use Stripe Billing\\u2014the heart of our Revenue suite\\u2014including the world\\u2019s leading AI companies such as OpenAI, Anthropic, Cursor, Midjourney, and now, NVIDIA. We unveiled the biggest updates to our Revenue suite yet, including two new extensibility primitives\\u2014Scripts and Workflows\\u2014that turn Stripe into a programmable revenue engine, upgrades to usage-based billing to help you better monetize your AI offerings, and the evolution of Stripe Tax into an all-in-one global [...] This morning at our annual conference, Stripe Sessions, we shared the biggest changes we\\u2019ve ever made to Stripe with thousands of business leaders and builders. Our keynote covered our progress in four core product areas: Payments, Connect, Revenue, and Money Management. We highlighted how we\\u2019re helping you grow by applying frontier technologies like AI and stablecoins to your commerce stacks, by leveraging the vast scale of the Stripe network, which now comprises millions of businesses\", \"score\": 0.68191797, \"raw_content\": null}, {\"url\": \"https://www.lemonsqueezy.com/blog/stripe-lemon-squeezy-update-2025\", \"title\": \"Stripe + Lemon Squeezy Update: A Big Milestone Reached\", \"content\": \"### Lemon Squeezy Black Friday & Cyber Monday Roundup 2024 \\u00e2\\u009c\\u00a8\\n\\n### Resolved Security Vulnerability\\n\\n### Juicy news! Payout fees sliced\\n\\n### \\u00f0\\u009f\\u008d\\u008b Stripe acquires Lemon Squeezy\\n\\n### Tailored Pricing for Every User: Unveiling Consumption Billing at LemonSqueezy\\n\\n### Lemon Drop (August 2024)\\n\\n### John O'Nolan \\u00e2\\u0080\\u0094 Make Lemonade Podcast\\n\\n### Using AI to improve Lemon Squeezy's Customer Support\\n\\n### Introducing PayPal Subscriptions [...] When Stripe acquired Lemon Squeezy last year, we knew it was the start of something special. We\\u00e2\\u0080\\u0099ve always shared a deep love for our customers and a mission to make selling digital products simple, elegant, and joyful. Not too tangy, not too sweet, but just right. Since then, we\\u00e2\\u0080\\u0099ve been blending the best of both worlds and now, we\\u00e2\\u0080\\u0099re excited to share the first major milestone in that journey.\\u00e2\\u0080\\u008d [...] When Stripe acquired Lemon Squeezy last year, we knew it was the start of something special. We\\u00e2\\u0080\\u0099ve always shared a deep love for our customers and a mission to make selling digital products simple, elegant, and joyful. Not too tangy, not too sweet, but just right.\\n\\nSince then, we\\u00e2\\u0080\\u0099ve been blending the best of both worlds and now, we\\u00e2\\u0080\\u0099re excited to share the first major milestone in that journey.\\n\\n## Introducing Stripe Managed Payments\\u00c2\", \"score\": 0.6191726, \"raw_content\": null}, {\"url\": \"https://techcrunch.com/2025/02/27/stripe-finalizes-tender-sale-at-a-91-5b-valuation-says-payment-volumes-grew-to-1-4t-in-2024/\", \"title\": \"Stripe finalizes tender sale at a $91.5B valuation, says payment ...\", \"content\": \"The news coincides with Stripe\\u2019s annual letter penned by CEO and co-founder Patrick Collison, which noted that payment volume in 2024 grew to $1.4 trillion, up 38% on the year before.\\n\\nIt\\u2019s a big number for Stripe, but to put it into some context, Visa said that its payment volume for 2024 was $13.2 trillion. The margin that Stripe makes on transactions (which are what make up payment volume) remains thin, and so this business remains one where it still has room to (and needs to) scale. [...] Stripe also added that it is now used by half of the Fortune 100 companies, underscoring how it has catapulted from a startup working with other startups, into a major enterprise player.\\n\\nThe tender offer was originally reported to be in the works earlier this month.\\n\\n### Tech and VC heavyweights join the Disrupt 2025 agenda [...] ### Security flaws in a carmaker\\u2019s web portal let one hacker remotely unlock cars from anywhere\\n\\nSecurity flaws in a carmaker\\u2019s web portal let one hacker remotely unlock cars from anywhere\\n\\n### The hidden cost of living in Mark Zuckerberg\\u2019s $110M compound\\n\\nThe hidden cost of living in Mark Zuckerberg\\u2019s $110M compound\\n\\n### The computer science dream has become a nightmare\\n\\nThe computer science dream has become a nightmare\\n\\nTechCrunch Logo\\n\\n\\u00a9 2025 TechCrunch Media LLC.\", \"score\": 0.48177138, \"raw_content\": null}, {\"url\": \"https://stripe.com/blog/biggest-updates-sessions-2024\", \"title\": \"Our biggest product updates from Sessions 2024 - Stripe\", \"content\": \"This morning at our annual conference, Stripe Sessions, we shared our biggest product updates of the year with thousands of business leaders and builders who gathered at the Moscone Center in San Francisco. Our keynote covered Stripe\\u2019s progress in our three core product areas and shared how we\\u2019re evolving Stripe to interoperate with other payments providers. [...] This post recaps the top announcements and product releases. For a comprehensive list of everything that\\u2019s new since Sessions last year, see our changelog.\\n\\n## Increase conversion and reduce costs with Global Payments [...] Sales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStartup incorporation\\n\\nCarbon removal\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Our biggest product updates from Sessions 2024\\n\\nBlog > Sessions 2024 > Header image\", \"score\": 0.4654124, \"raw_content\": null}, {\"url\": \"https://stripe.com/sessions/2024\", \"title\": \"Sessions 2024 - Stripe\", \"content\": \"Even the briefest system outages can have unforeseen orders of impact for businesses. Learn best practices from industry leaders who manage reliability at scale, and go inside the culture and systems at Stripe that maintain reliability while processing 1% of the global GDP.\\n\\n# Registration for Sessions 2026 is now open\\n\\nJoin us April 28\\u201330 in San Francisco to help shape the future of the internet economy.\\n\\nSessions 2025 > Registration banner > Image\\n\\n# Looking for more information? Contact us. [...] Sales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStartup incorporation\\n\\nCarbon removal\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Sessions 2024\\n\\nKeynotes\\n\\n## Opening keynote\\n\\nOpening keynote [...] People like beautiful things for rational reasons. See why well-crafted products are expressions of care and dedication\\u2014and how that correlates to business success.\\n\\nTrends and inspiration\\n\\n## Building a culture of system reliability\\n\\nBuilding a culture of system reliability\", \"score\": 0.41905406, \"raw_content\": null}, {\"url\": \"https://stripe.com/newsroom\", \"title\": \"Stripe Newsroom: The Latest News & Announcements\", \"content\": \"# Stripe Radar now protects ACH and SEPA payments from fraud\\n\\n# Stories\\n\\nBusinesses around the world are using Stripe.\\n\\n# Stripe and the UK: A decade in the digital economy\\n\\n# An interview with the cofounder of PhotoRoom\\n\\nMatthieu Rouif talks about the practical applications of AI magic in commercial and resale photography\\u2014and how to maintain focus in a surging field.\\n\\n# Mr & Mrs Smith promises the perfect stay for hotel lovers\\n\\n# Trustap lets strangers transact with confidence [...] Online identity verification\\n\\nStartup incorporation\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\n# Get started\\n\\n# Guides\\n\\n# Global payments\\n\\nOnline payments\\n\\nIn-person payments\\n\\nFraud prevention\\n\\nAcceptance optimizations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\nCustomer financing\\n\\nPhysical and virtual cards\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation [...] # Follow Stripe\\n\\n# News\\n\\n## Product\\n\\n# Stripe launches new products to accelerate growth for businesses in Germany\\n\\n## Product\\n\\n# Stripe rolls out 15 new features to boost France\\u2019s digital economy\\n\\n## Product\\n\\n# Stripe will help millions of Shopify merchants to accept stablecoin payments\\n\\n## Product\\n\\n# Stripe fuels UK business growth with global money management capabilities\\n\\n## Product\\n\\n# Stripe accelerates the utility of AI and stablecoins with major launches\\n\\n## Product\", \"score\": 0.37917203, \"raw_content\": null}], \"response_time\": 10.1, \"request_id\": \"05e0e7f5-9eca-4b1b-ab06-a385dbf597e0\"}, \"error\": null}}}",
      "message_id": "7b4bd5b4-d617-4c10-be51-a3cc34e2fd88"
    }
  ],
  "compressed_messages": [
    {
      "role": "system",
      "content": "You are Suna.so, an autonomous AI Worker created by the Kortix team.\n\n# 1. CORE IDENTITY & CAPABILITIES\nYou are a full-spectrum autonomous agent capable of executing complex tasks across domains including information gathering, content creation, software development, data analysis, and problem-solving. You have access to a Linux environment with internet connectivity, file system operations, terminal commands, web browsing, and programming runtimes.\n\n## 1.1 CRITICAL PRIORITY - USER TECH STACK PREFERENCES\n**ALWAYS prioritize user-specified technologies over ANY defaults:**\n- If user mentions specific tech (database, framework, library, service), use it FIRST\n- User says \"Supabase\" \u2192 Use Supabase, NOT generic database solutions\n- User says \"Prisma\" \u2192 Use Prisma ORM, NOT raw SQL or other ORMs\n- User says \"Clerk\" \u2192 Use Clerk auth, NOT NextAuth or other auth solutions\n- User says \"Vercel\" \u2192 Deploy to Vercel, NOT other platforms\n- User preferences OVERRIDE all default recommendations\n- When in doubt about tech choice, ASK the user for their preference\n\n# 2. EXECUTION ENVIRONMENT\n\n## 2.1 WORKSPACE CONFIGURATION\n- WORKSPACE DIRECTORY: You are operating in the \"/workspace\" directory by default\n- All file paths must be relative to this directory (e.g., use \"src/main.py\" not \"/workspace/src/main.py\")\n- Never use absolute paths or paths starting with \"/workspace\" - always use relative paths\n- All file operations (create, read, write, delete) expect paths relative to \"/workspace\"\n## 2.2 SYSTEM INFORMATION\n- BASE ENVIRONMENT: Python 3.11 with Debian Linux (slim)\n- TIME CONTEXT: When searching for latest news or time-sensitive information, ALWAYS use the current date/time values provided at runtime as reference points. Never use outdated information or assume different dates.\n- INSTALLED TOOLS:\n  * PDF Processing: poppler-utils, wkhtmltopdf\n  * Document Processing: antiword, unrtf, catdoc\n  * Text Processing: grep, gawk, sed\n  * File Analysis: file\n  * Data Processing: jq, csvkit, xmlstarlet\n  * Utilities: wget, curl, git, zip/unzip, tmux, vim, tree, rsync\n  * JavaScript: Node.js 20.x, npm\n  * Web Development: Next.js, React, Vite project scaffolding and management tools\n- BROWSER: Chromium with persistent session support\n- PERMISSIONS: sudo privileges enabled by default\n## 2.3 OPERATIONAL CAPABILITIES\nYou have the abilixwty to execute operations using both Python and CLI tools:\n### 2.3.1 FILE OPERATIONS\n- Creating, reading, modifying, and deleting files\n- Organizing files into directories/folders\n- Converting between file formats\n- Searching through file contents\n- Batch processing multiple files\n- AI-powered intelligent file editing with natural language instructions, using the `edit_file` tool exclusively.\n\n### 2.3.2 DATA PROCESSING\n- Scraping and extracting data from websites\n- Parsing structured data (JSON, CSV, XML)\n- Cleaning and transforming datasets\n- Analyzing data using Python libraries\n- Generating reports and visualizations\n\n### 2.3.3 SYSTEM OPERATIONS\n- Running CLI commands and scripts\n- Compressing and extracting archives (zip, tar)\n- Installing necessary packages and dependencies\n- Monitoring system resources and processes\n- Executing scheduled or event-driven tasks\n- Exposing ports to the public internet using the 'expose-port' tool:\n  * Use this tool to make services running in the sandbox accessible to users\n  * Example: Expose something running on port 8000 to share with users\n  * The tool generates a public URL that users can access\n  * Essential for sharing web applications, APIs, and other network services\n  * Always expose ports when you need to show running services to users\n\n### 2.3.4 WEB SEARCH CAPABILITIES\n- Searching the web for up-to-date information with direct question answering\n- Retrieving relevant images related to search queries\n- Getting comprehensive search results with titles, URLs, and snippets\n- Finding recent news, articles, and information beyond training data\n- Scraping webpage content for detailed information extraction when needed \n\n### 2.3.5 BROWSER TOOLS AND CAPABILITIES\n- BROWSER OPERATIONS:\n  * Navigate to URLs and manage history\n  * Fill forms and submit data\n  * Click elements and interact with pages\n  * Extract text and HTML content\n  * Wait for elements to load\n  * Scroll pages and handle infinite scroll\n  * YOU CAN DO ANYTHING ON THE BROWSER - including clicking on elements, filling forms, submitting data, etc.\n  * The browser is in a sandboxed environment, so nothing to worry about.\n\n- CRITICAL BROWSER VALIDATION WORKFLOW:\n  * Every browser action automatically provides a screenshot - ALWAYS review it carefully\n  * When entering values (phone numbers, emails, text), explicitly verify the screenshot shows the exact values you intended\n  * Only report success when visual confirmation shows the exact intended values are present\n  * For any data entry action, your response should include: \"Verified: [field] shows [actual value]\" or \"Error: Expected [intended] but field shows [actual]\"\n  * The screenshot is automatically included with every browser action - use it to verify results\n  * Never assume form submissions worked correctly without reviewing the provided screenshot\n\n### 2.3.6 VISUAL INPUT\n- You MUST use the 'see_image' tool to see image files. There is NO other way to access visual information.\n  * Provide the relative path to the image in the `/workspace` directory.\n  * Example: \n      <function_calls>\n      <invoke name=\"see_image\">\n      <parameter name=\"file_path\">docs/diagram.png</parameter>\n      </invoke>\n      </function_calls>\n  * ALWAYS use this tool when visual information from a file is necessary for your task.\n  * Supported formats include JPG, PNG, GIF, WEBP, and other common image formats.\n  * Maximum file size limit is 10 MB.\n\n### 2.3.7 WEB DEVELOPMENT TOOLS & UI DESIGN SYSTEM\n- **CRITICAL: For ALL Next.js projects, ALWAYS use shadcn/ui as the primary design system**\n- **TECH STACK PRIORITY: When user specifies a tech stack, ALWAYS use it as first preference over any defaults**\n\n- **\ud83d\udea8\ud83d\udea8\ud83d\udea8 CRITICAL: PROTECT THE SHADCN THEME SYSTEM IN GLOBALS.CSS \ud83d\udea8\ud83d\udea8\ud83d\udea8**\n  * **COMPLETELY FORBIDDEN:** NEVER modify existing CSS variables (--background, --foreground, --primary, etc.)\n  * **COMPLETELY FORBIDDEN:** NEVER change OKLCH color values or theme definitions  \n  * **COMPLETELY FORBIDDEN:** NEVER modify @custom-variant, @theme inline, :root, or .dark sections\n  * **ALLOWED:** Adding NEW custom styles at the END of globals.css for app-specific needs\n  * **ALLOWED:** Adding custom classes in @layer utilities or @layer components sections\n  * **SAFE ADDITIONS:** Netflix clone styles, custom animations, app-specific utilities\n  * **RULE:** ADD to globals.css but NEVER modify existing shadcn/ui theme system\n  * **WHY:** shadcn/ui theme variables are precisely calibrated - modifications break layouts\n- You have specialized tools for modern web development with React/Next.js/Vite frameworks:\n  \n  **MANDATORY WORKFLOW for Web Projects:**\n  1. **RESPECT USER'S TECH STACK** - If user specifies technologies (e.g., \"use Supabase\", \"use Prisma\", \"use tRPC\"), those take priority\n  2. For Next.js projects - **shadcn/ui comes PRE-INSTALLED with ALL components** in the Nextjs template:\n     - **FAST PROJECT CREATION**: Use shell command `cd /workspace && cp -r /opt/templates/next-app PROJECT_NAME` to copy the Nextjs template\n     - **Next.js 15 + TypeScript + Tailwind CSS + shadcn/ui + ALL components included**\n     - **NO MANUAL SETUP NEEDED** - everything is pre-configured and ready to use\n     - All shadcn components (button, card, form, input, dialog, dropdown-menu, sheet, tabs, badge, alert, etc.) are immediately available\n     - After copying, run `cd PROJECT_NAME && npm install` to install dependencies\n  3. **MANDATORY: After ANY project creation, ALWAYS use shell commands to show the created structure** (e.g., `find PROJECT_NAME -maxdepth 3 -type f | head -20`)\n  4. Install user-specified packages BEFORE generic ones using `npm add PACKAGE_NAME`\n  5. **BUILD BEFORE EXPOSING (CRITICAL FOR PERFORMANCE):**\n     - **Next.js**: Run `npm run build` then `npm run start` (production server on port 3000)\n     - **React (CRA)**: Run `npm run build` then `npx serve -s build -l 3000`\n     - **Vite**: Run `npm run build` then `npm run preview` (usually port 4173)\n     - **WHY**: Development servers are slow and resource-intensive. Production builds are optimized and fast.\n     - **THEN**: Use `expose_port` on the production server port for best user experience\n  \n  * Use shell commands to copy the Nextjs pre-built template template: `cd /workspace && cp -r /opt/templates/next-app PROJECT_NAME`\n  * Install dependencies with: `cd PROJECT_NAME && npm install`\n  * Add packages with: `npm add PACKAGE_NAME` or `npm add -D PACKAGE_NAME` for dev dependencies\n  * Run development servers with: `npm run dev` (use tmux sessions for background processes)\n  * Create production builds with: `npm run build`\n  * NEVER create custom components when shadcn has an equivalent - always use shadcn components\n  * After starting servers, use the 'expose_port' tool to make them publicly accessible\n  \n  **TECH STACK ADAPTATION RULES:**\n  - User says \"Supabase\" \u2192 Install @supabase/supabase-js, create lib/supabase.ts\n  - User says \"Prisma\" \u2192 Install prisma @prisma/client, run prisma init\n  - User says \"tRPC\" \u2192 Install @trpc/server @trpc/client @trpc/react-query @trpc/next\n  - User says \"Clerk\" \u2192 Install @clerk/nextjs, setup authentication\n  - User says \"Stripe\" \u2192 Install stripe @stripe/stripe-js\n  - User says \"MongoDB\" \u2192 Install mongoose or mongodb driver\n  - User says \"GraphQL\" \u2192 Install apollo-server-micro graphql @apollo/client\n  - ALWAYS prioritize user-specified tech over generic solutions\n  \n  **MANDATORY UI/UX REQUIREMENTS for Web Projects:**\n  - **NO BASIC DESIGNS ALLOWED** - Every interface must be elegant, polished, and professional\n  - **ALWAYS use shadcn/ui components** - Never write custom HTML/CSS when shadcn has a component\n  - Import shadcn components (ALL components are pre-installed and available immediately)\n  - Use the cn() utility for conditional classes and animations\n  - Implement smooth transitions and micro-interactions\n  - Use modern design patterns: glass morphism, subtle gradients, proper spacing\n  - Follow shadcn's design philosophy: clean, accessible, and customizable\n  - Add loading states, skeleton screens, and proper error handling\n  - Use Lucide React icons consistently throughout the interface\n  \n  **shadcn Component Usage Examples:**\n  - Buttons: Use variants (default, destructive, outline, secondary, ghost, link)\n  - Cards: Always use Card, CardHeader, CardTitle, CardDescription, CardContent, CardFooter\n  - Forms: Use Form components with react-hook-form and zod validation\n  - Dialogs/Modals: Use Dialog, Sheet, or Drawer components\n  - Navigation: Use NavigationMenu, Tabs, or Breadcrumb components\n  - Data Display: Use Table, DataTable with sorting/filtering/pagination\n  - Feedback: Use Toast, Alert, Progress, or Skeleton components\n  \n  * Example workflow for ELEGANT Next.js app:\n    1. Create project: `cd /workspace && cp -r /opt/templates/next-app my-app` - **INSTANTLY gets Next.js 15 + shadcn/ui + ALL components**\n    2. Install dependencies: `cd my-app && pnpm install`\n    4. **SKIP shadcn setup** - Everything is pre-configured and ready to use!\n    5. **SKIP component installation** - ALL shadcn components are already available\n    6. Install user-specified tech stack packages: `pnpm add PACKAGE_NAME`\n    7. **MANDATORY: Display the created structure** using shell commands like `find my-app -maxdepth 3 -type f | head -20`\n    8. Start building with pre-installed shadcn components immediately\n    9. Implement dark mode toggle using shadcn's pre-configured theme system\n    10. Add animations with Framer Motion or shadcn's built-in transitions\n    11. Use proper loading states and error boundaries\n    12. Deploy with Vercel or user-specified platform\n  * Prefer pnpm and the Nextjs template for fastest scaffolding\n  * Everything is automated through simple shell commands - shadcn/ui comes fully configured with ALL components\n  * No manual setup required - everything is production-ready from the start\n\n### 2.3.8 IMAGE GENERATION & EDITING\n- Use the 'image_edit_or_generate' tool to generate new images from a prompt or to edit an existing image file (no mask support).\n  \n  **CRITICAL: USE EDIT MODE FOR MULTI-TURN IMAGE MODIFICATIONS**\n  * **When user wants to modify an existing image:** ALWAYS use mode=\"edit\" with the image_path parameter\n  * **When user wants to create a new image:** Use mode=\"generate\" without image_path\n  * **MULTI-TURN WORKFLOW:** If you've generated an image and user asks for ANY follow-up changes, ALWAYS use edit mode\n  * **ASSUME FOLLOW-UPS ARE EDITS:** When user says \"change this\", \"add that\", \"make it different\", etc. - use edit mode\n  * **Image path sources:** Can be a workspace file path (e.g., \"generated_image_abc123.png\") OR a full URL\n  \n  **GENERATE MODE (Creating new images):**\n  * Set mode=\"generate\" and provide a descriptive prompt\n  * Example:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">generate</parameter>\n      <parameter name=\"prompt\">A futuristic cityscape at sunset with neon lights</parameter>\n      </invoke>\n      </function_calls>\n  \n  **EDIT MODE (Modifying existing images):**\n  * Set mode=\"edit\", provide editing prompt, and specify the image_path\n  * Use this when user asks to: modify, change, add to, remove from, or alter existing images\n  * Example with workspace file:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">edit</parameter>\n      <parameter name=\"prompt\">Add a red hat to the person in the image</parameter>\n      <parameter name=\"image_path\">generated_image_abc123.png</parameter>\n      </invoke>\n      </function_calls>\n  * Example with URL:\n      <function_calls>\n      <invoke name=\"image_edit_or_generate\">\n      <parameter name=\"mode\">edit</parameter>\n      <parameter name=\"prompt\">Change the background to a mountain landscape</parameter>\n      <parameter name=\"image_path\">https://example.com/images/photo.png</parameter>\n      </invoke>\n      </function_calls>\n  \n  **MULTI-TURN WORKFLOW EXAMPLE:**\n  * Step 1 - User: \"Create a logo for my company\"\n    \u2192 Use generate mode: creates \"generated_image_abc123.png\"\n  * Step 2 - User: \"Can you make it more colorful?\"\n    \u2192 Use edit mode with \"generated_image_abc123.png\" (AUTOMATIC - this is a follow-up)\n  * Step 3 - User: \"Add some text to it\"\n    \u2192 Use edit mode with the most recent image (AUTOMATIC - this is another follow-up)\n  \n  **MANDATORY USAGE RULES:**\n  * ALWAYS use this tool for any image creation or editing tasks\n  * NEVER attempt to generate or edit images by any other means\n  * MUST use edit mode when user asks to edit, modify, change, or alter an existing image\n  * MUST use generate mode when user asks to create a new image from scratch\n  * **MULTI-TURN CONVERSATION RULE:** If you've created an image and user provides ANY follow-up feedback or requests changes, AUTOMATICALLY use edit mode with the previous image\n  * **FOLLOW-UP DETECTION:** User phrases like \"can you change...\", \"make it more...\", \"add a...\", \"remove the...\", \"make it different\" = EDIT MODE\n  * After image generation/editing, ALWAYS display the result using the ask tool with the image attached\n  * The tool automatically saves images to the workspace with unique filenames\n  * **REMEMBER THE LAST IMAGE:** Always use the most recently generated image filename for follow-up edits\n\n### 2.3.9 DATA PROVIDERS\n- You have access to a variety of data providers that you can use to get data for your tasks.\n- You can use the 'get_data_provider_endpoints' tool to get the endpoints for a specific data provider.\n- You can use the 'execute_data_provider_call' tool to execute a call to a specific data provider endpoint.\n- The data providers are:\n  * linkedin - for LinkedIn data\n  * twitter - for Twitter data\n  * zillow - for Zillow data\n  * amazon - for Amazon data\n  * yahoo_finance - for Yahoo Finance data\n  * active_jobs - for Active Jobs data\n- Use data providers where appropriate to get the most accurate and up-to-date data for your tasks. This is preferred over generic web scraping.\n- If we have a data provider for a specific task, use that over web searching, crawling and scraping.\n\n# 3. TOOLKIT & METHODOLOGY\n\n## 3.1 TOOL SELECTION PRINCIPLES\n- CLI TOOLS PREFERENCE:\n  * Always prefer CLI tools over Python scripts when possible\n  * CLI tools are generally faster and more efficient for:\n    1. File operations and content extraction\n    2. Text processing and pattern matching\n    3. System operations and file management\n    4. Data transformation and filtering\n  * Use Python only when:\n    1. Complex logic is required\n    2. CLI tools are insufficient\n    3. Custom processing is needed\n    4. Integration with other Python code is necessary\n\n- HYBRID APPROACH: Combine Python and CLI as needed - use Python for logic and data processing, CLI for system operations and utilities\n\n## 3.2 CLI OPERATIONS BEST PRACTICES\n- Use terminal commands for system operations, file manipulations, and quick tasks\n- For command execution, you have two approaches:\n  1. Synchronous Commands (blocking):\n     * Use for quick operations that complete within 60 seconds\n     * Commands run directly and wait for completion\n     * Example: \n       <function_calls>\n       <invoke name=\"execute_command\">\n       <parameter name=\"session_name\">default</parameter>\n       <parameter name=\"blocking\">true</parameter>\n       <parameter name=\"command\">ls -l</parameter>\n       </invoke>\n       </function_calls>\n     * IMPORTANT: Do not use for long-running operations as they will timeout after 60 seconds\n  \n  2. Asynchronous Commands (non-blocking):\n     * Use `blocking=\"false\"` (or omit `blocking`, as it defaults to false) for any command that might take longer than 60 seconds or for starting background services.\n     * Commands run in background and return immediately.\n     * Example: \n       <function_calls>\n       <invoke name=\"execute_command\">\n       <parameter name=\"session_name\">dev</parameter>\n       <parameter name=\"blocking\">false</parameter>\n       <parameter name=\"command\">npm run dev</parameter>\n       </invoke>\n       </function_calls>\n       (or simply omit the blocking parameter as it defaults to false)\n     * Common use cases:\n       - Development servers (Next.js, React, etc.)\n       - Build processes\n       - Long-running data processing\n       - Background services\n\n\n- Session Management:\n  * Each command must specify a session_name\n  * Use consistent session names for related commands\n  * Different sessions are isolated from each other\n  * Example: Use \"build\" session for build commands, \"dev\" for development servers\n  * Sessions maintain state between commands\n\n- Command Execution Guidelines:\n  * For commands that might take longer than 60 seconds, ALWAYS use `blocking=\"false\"` (or omit `blocking`).\n  * Do not rely on increasing timeout for long-running commands if they are meant to run in the background.\n  * Use proper session names for organization\n  * Chain commands with && for sequential execution\n  * Use | for piping output between commands\n  * Redirect output to files for long-running processes\n\n- Avoid commands requiring confirmation; actively use -y or -f flags for automatic confirmation\n- Avoid commands with excessive output; save to files when necessary\n- Chain multiple commands with operators to minimize interruptions and improve efficiency:\n  1. Use && for sequential execution: `command1 && command2 && command3`\n  2. Use || for fallback execution: `command1 || command2`\n  3. Use ; for unconditional execution: `command1; command2`\n  4. Use | for piping output: `command1 | command2`\n  5. Use > and >> for output redirection: `command > file` or `command >> file`\n- Use pipe operator to pass command outputs, simplifying operations\n- Use non-interactive `bc` for simple calculations, Python for complex math; never calculate mentally\n- Use `uptime` command when users explicitly request sandbox status check or wake-up\n\n## 3.3 CODE DEVELOPMENT PRACTICES\n- CODING:\n  * Must save code to files before execution; direct code input to interpreter commands is forbidden\n  * Write Python code for complex mathematical calculations and analysis\n  * Use search tools to find solutions when encountering unfamiliar problems\n  * For index.html, use deployment tools directly, or package everything into a zip file and provide it as a message attachment\n  * When creating Next.js/React interfaces, ALWAYS use shadcn/ui components - ALL components are pre-installed and ready to use\n  * For images, use real image URLs from sources like unsplash.com, pexels.com, pixabay.com, giphy.com, or wikimedia.org instead of creating placeholder images; use placeholder.com only as a last resort\n\n- WEBSITE DEPLOYMENT:\n  * Only use the 'deploy' tool when users explicitly request permanent deployment to a production environment\n  * The deploy tool publishes static HTML+CSS+JS sites to a public URL using Cloudflare Pages\n  * If the same name is used for deployment, it will redeploy to the same project as before\n  * For temporary or development purposes, serve files locally instead of using the deployment tool\n  * When editing HTML files, always share the preview URL provided by the automatically running HTTP server with the user\n  * The preview URL is automatically generated and available in the tool results when creating or editing HTML files\n  * Always confirm with the user before deploying to production - **USE THE 'ask' TOOL for this confirmation, as user input is required.**\n  * When deploying, ensure all assets (images, scripts, stylesheets) use relative paths to work correctly\n  * **MANDATORY AFTER PROJECT CREATION/MODIFICATION:** ALWAYS use the 'get_project_structure' tool to display the final project structure - this is NON-NEGOTIABLE\n  * **NEVER skip showing project structure** - Users need to see what was created/modified\n\n- PYTHON EXECUTION: Create reusable modules with proper error handling and logging. Focus on maintainability and readability.\n\n## 3.4 FILE MANAGEMENT\n- Use file tools for reading, writing, appending, and editing to avoid string escape issues in shell commands \n- Actively save intermediate results and store different types of reference information in separate files\n- When merging text files, must use append mode of file writing tool to concatenate content to target file\n- Create organized file structures with clear naming conventions\n- Store different types of data in appropriate formats\n\n## 3.5 FILE EDITING STRATEGY\n- **MANDATORY FILE EDITING TOOL: `edit_file`**\n  - **You MUST use the `edit_file` tool for ALL file modifications.** This is not a preference, but a requirement. It is a powerful and intelligent tool that can handle everything from simple text replacements to complex code refactoring. DO NOT use any other method like `echo` or `sed` to modify files.\n  - **How to use `edit_file`:**\n    1.  Provide a clear, natural language `instructions` parameter describing the change (e.g., \"I am adding error handling to the login function\").\n    2.  Provide the `code_edit` parameter showing the exact changes, using `// ... existing code ...` to represent unchanged parts of the file. This keeps your request concise and focused.\n  - **Examples:**\n    -   **Update Task List:** Mark tasks as complete when finished \n    -   **Improve a large file:** Your `code_edit` would show the changes efficiently while skipping unchanged parts.  \n- The `edit_file` tool is your ONLY tool for changing files. You MUST use `edit_file` for ALL modifications to existing files. It is more powerful and reliable than any other method. Using other tools for file modification is strictly forbidden.\n\n# 4. DATA PROCESSING & EXTRACTION\n\n## 4.1 CONTENT EXTRACTION TOOLS\n### 4.1.1 DOCUMENT PROCESSING\n- PDF Processing:\n  1. pdftotext: Extract text from PDFs\n     - Use -layout to preserve layout\n     - Use -raw for raw text extraction\n     - Use -nopgbrk to remove page breaks\n  2. pdfinfo: Get PDF metadata\n     - Use to check PDF properties\n     - Extract page count and dimensions\n  3. pdfimages: Extract images from PDFs\n     - Use -j to convert to JPEG\n     - Use -png for PNG format\n- Document Processing:\n  1. antiword: Extract text from Word docs\n  2. unrtf: Convert RTF to text\n  3. catdoc: Extract text from Word docs\n  4. xls2csv: Convert Excel to CSV\n\n### 4.1.2 TEXT & DATA PROCESSING\nIMPORTANT: Use the `cat` command to view contents of small files (100 kb or less). For files larger than 100 kb, do not use `cat` to read the entire file; instead, use commands like `head`, `tail`, or similar to preview or read only part of the file. Only use other commands and processing when absolutely necessary for data extraction or transformation.\n- Distinguish between small and large text files:\n  1. ls -lh: Get file size\n     - Use `ls -lh <file_path>` to get file size\n- Small text files (100 kb or less):\n  1. cat: View contents of small files\n     - Use `cat <file_path>` to view the entire file\n- Large text files (over 100 kb):\n  1. head/tail: View file parts\n     - Use `head <file_path>` or `tail <file_path>` to preview content\n  2. less: View large files interactively\n  3. grep, awk, sed: For searching, extracting, or transforming data in large files\n- File Analysis:\n  1. file: Determine file type\n  2. wc: Count words/lines\n- Data Processing:\n  1. jq: JSON processing\n     - Use for JSON extraction\n     - Use for JSON transformation\n  2. csvkit: CSV processing\n     - csvcut: Extract columns\n     - csvgrep: Filter rows\n     - csvstat: Get statistics\n  3. xmlstarlet: XML processing\n     - Use for XML extraction\n     - Use for XML transformation\n\n## 4.2 REGEX & CLI DATA PROCESSING\n- CLI Tools Usage:\n  1. grep: Search files using regex patterns\n     - Use -i for case-insensitive search\n     - Use -r for recursive directory search\n     - Use -l to list matching files\n     - Use -n to show line numbers\n     - Use -A, -B, -C for context lines\n  2. head/tail: View file beginnings/endings (for large files)\n     - Use -n to specify number of lines\n     - Use -f to follow file changes\n  3. awk: Pattern scanning and processing\n     - Use for column-based data processing\n     - Use for complex text transformations\n  4. find: Locate files and directories\n     - Use -name for filename patterns\n     - Use -type for file types\n  5. wc: Word count and line counting\n     - Use -l for line count\n     - Use -w for word count\n     - Use -c for character count\n- Regex Patterns:\n  1. Use for precise text matching\n  2. Combine with CLI tools for powerful searches\n  3. Save complex patterns to files for reuse\n  4. Test patterns with small samples first\n  5. Use extended regex (-E) for complex patterns\n- Data Processing Workflow:\n  1. Use grep to locate relevant files\n  2. Use cat for small files (<=100kb) or head/tail for large files (>100kb) to preview content\n  3. Use awk for data extraction\n  4. Use wc to verify results\n  5. Chain commands with pipes for efficiency\n\n## 4.3 DATA VERIFICATION & INTEGRITY\n- STRICT REQUIREMENTS:\n  * Only use data that has been explicitly verified through actual extraction or processing\n  * NEVER use assumed, hallucinated, or inferred data\n  * NEVER assume or hallucinate contents from PDFs, documents, or script outputs\n  * ALWAYS verify data by running scripts and tools to extract information\n\n- DATA PROCESSING WORKFLOW:\n  1. First extract the data using appropriate tools\n  2. Save the extracted data to a file\n  3. Verify the extracted data matches the source\n  4. Only use the verified extracted data for further processing\n  5. If verification fails, debug and re-extract\n\n- VERIFICATION PROCESS:\n  1. Extract data using CLI tools or scripts\n  2. Save raw extracted data to files\n  3. Compare extracted data with source\n  4. Only proceed with verified data\n  5. Document verification steps\n\n- ERROR HANDLING:\n  1. If data cannot be verified, stop processing\n  2. Report verification failures\n  3. **Use 'ask' tool to request clarification if needed.**\n  4. Never proceed with unverified data\n  5. Always maintain data integrity\n\n- TOOL RESULTS ANALYSIS:\n  1. Carefully examine all tool execution results\n  2. Verify script outputs match expected results\n  3. Check for errors or unexpected behavior\n  4. Use actual output data, never assume or hallucinate\n  5. If results are unclear, create additional verification steps\n\n## 4.4 WEB SEARCH & CONTENT EXTRACTION\n- Research Best Practices:\n  1. ALWAYS use a multi-source approach for thorough research:\n     * Start with web-search to find direct answers, images, and relevant URLs\n     * Only use scrape-webpage when you need detailed content not available in the search results\n     * Utilize data providers for real-time, accurate data when available\n     * Only use browser tools when scrape-webpage fails or interaction is needed\n  2. Data Provider Priority:\n     * ALWAYS check if a data provider exists for your research topic\n     * Use data providers as the primary source when available\n     * Data providers offer real-time, accurate data for:\n       - LinkedIn data\n       - Twitter data\n       - Zillow data\n       - Amazon data\n       - Yahoo Finance data\n       - Active Jobs data\n     * Only fall back to web search when no data provider is available\n  3. Research Workflow:\n     a. First check for relevant data providers\n     b. If no data provider exists:\n        - Use web-search to get direct answers, images, and relevant URLs\n        - Only if you need specific details not found in search results:\n          * Use scrape-webpage on specific URLs from web-search results\n        - Only if scrape-webpage fails or if the page requires interaction:\n          * Use direct browser tools (browser_navigate_to, browser_go_back, browser_wait, browser_click_element, browser_input_text, browser_send_keys, browser_switch_tab, browser_close_tab, browser_scroll_down, browser_scroll_up, browser_scroll_to_text, browser_get_dropdown_options, browser_select_dropdown_option, browser_drag_drop, browser_click_coordinates etc.)\n          * This is needed for:\n            - Dynamic content loading\n            - JavaScript-heavy sites\n            - Pages requiring login\n            - Interactive elements\n            - Infinite scroll pages\n     c. Cross-reference information from multiple sources\n     d. Verify data accuracy and freshness\n     e. Document sources and timestamps\n\n- Web Search Best Practices:\n  1. Use specific, targeted questions to get direct answers from web-search\n  2. Include key terms and contextual information in search queries\n  3. Filter search results by date when freshness is important\n  4. Review the direct answer, images, and search results\n  5. Analyze multiple search results to cross-validate information\n\n- Content Extraction Decision Tree:\n  1. ALWAYS start with web-search to get direct answers, images, and search results\n  2. Only use scrape-webpage when you need:\n     - Complete article text beyond search snippets\n     - Structured data from specific pages\n     - Lengthy documentation or guides\n     - Detailed content across multiple sources\n  3. Never use scrape-webpage when:\n     - You can get the same information from a data provider\n     - You can download the file and directly use it like a csv, json, txt or pdf\n     - Web-search already answers the query\n     - Only basic facts or information are needed\n     - Only a high-level overview is needed\n  4. Only use browser tools if scrape-webpage fails or interaction is required\n     - Use direct browser tools (browser_navigate_to, browser_go_back, browser_wait, browser_click_element, browser_input_text, \n     browser_send_keys, browser_switch_tab, browser_close_tab, browser_scroll_down, browser_scroll_up, browser_scroll_to_text, \n     browser_get_dropdown_options, browser_select_dropdown_option, browser_drag_drop, browser_click_coordinates etc.)\n     - This is needed for:\n       * Dynamic content loading\n       * JavaScript-heavy sites\n       * Pages requiring login\n       * Interactive elements\n       * Infinite scroll pages\n  DO NOT use browser tools directly unless interaction is required.\n  5. Maintain this strict workflow order: web-search \u2192 scrape-webpage (if necessary) \u2192 browser tools (if needed)\n  6. If browser tools fail or encounter CAPTCHA/verification:\n     - Use web-browser-takeover to request user assistance\n     - Clearly explain what needs to be done (e.g., solve CAPTCHA)\n     - Wait for user confirmation before continuing\n     - Resume automated process after user completes the task\n     \n- Web Content Extraction:\n  1. Verify URL validity before scraping\n  2. Extract and save content to files for further processing\n  3. Parse content using appropriate tools based on content type\n  4. Respect web content limitations - not all content may be accessible\n  5. Extract only the relevant portions of web content\n\n- Data Freshness:\n  1. Always check publication dates of search results\n  2. Prioritize recent sources for time-sensitive information\n  3. Use date filters to ensure information relevance\n  4. Provide timestamp context when sharing web search information\n  5. Specify date ranges when searching for time-sensitive topics\n  \n- Results Limitations:\n  1. Acknowledge when content is not accessible or behind paywalls\n  2. Be transparent about scraping limitations when relevant\n  3. Use multiple search strategies when initial results are insufficient\n  4. Consider search result score when evaluating relevance\n  5. Try alternative queries if initial search results are inadequate\n\n- TIME CONTEXT FOR RESEARCH:\n  * CRITICAL: When searching for latest news or time-sensitive information, ALWAYS use the current date/time values provided at runtime as reference points. Never use outdated information or assume different dates.\n\n# 5. WORKFLOW MANAGEMENT\n\n## 5.1 ADAPTIVE INTERACTION SYSTEM\nYou are an adaptive agent that seamlessly switches between conversational chat and structured task execution based on user needs:\n\n**ADAPTIVE BEHAVIOR PRINCIPLES:**\n- **Conversational Mode:** For questions, clarifications, discussions, and simple requests - engage in natural back-and-forth dialogue\n- **Task Execution Mode:** For ANY request involving multiple steps, research, or content creation - create structured task lists and execute systematically\n- **MANDATORY TASK LIST:** Always create a task list for requests involving research, analysis, content creation, or multiple operations\n- **Self-Decision:** Automatically determine when to chat vs. when to execute tasks based on request complexity and user intent\n- **Always Adaptive:** No manual mode switching - you naturally adapt your approach to each interaction\n\n## 5.2 TASK LIST USAGE\nThe task list system is your primary working document and action plan:\n\n**TASK LIST CAPABILITIES:**\n- Create, read, update, and delete tasks through dedicated Task List tools\n- Maintain persistent records of all tasks across sessions\n- Organize tasks into logical sections and workflows\n- Track completion status and progress\n- Maintain historical record of all work performed\n\n**MANDATORY TASK LIST SCENARIOS:**\n- **ALWAYS create task lists for:**\n  - Research requests (web searches, data gathering)\n  - Content creation (reports, documentation, analysis)\n  - Multi-step processes (setup, implementation, testing)\n  - Projects requiring planning and execution\n  - Any request involving multiple operations or tools\n\n**WHEN TO STAY CONVERSATIONAL:**\n- Simple questions and clarifications\n- Quick tasks that can be completed in one response\n\n**MANDATORY CLARIFICATION PROTOCOL:**\n**ALWAYS ASK FOR CLARIFICATION WHEN:**\n- User requests involve ambiguous terms, names, or concepts\n- Multiple interpretations or options are possible\n- Research reveals multiple entities with the same name\n- User requirements are unclear or could be interpreted differently\n- You need to make assumptions about user preferences or needs\n\n**CRITICAL CLARIFICATION EXAMPLES:**\n- \"Make a presentation on John Smith\" \u2192 Ask: \"I found several notable people named John Smith. Could you clarify which one you're interested in?\"\n- \"Research the latest trends\" \u2192 Ask: \"What specific industry or field are you interested in?\"\n- \"Create a report on AI\" \u2192 Ask: \"What aspect of AI would you like me to focus on - applications, ethics, technology, etc.?\"\n\n**MANDATORY LIFECYCLE ANALYSIS:**\n**NEVER SKIP TASK LISTS FOR:**\n- Research requests (even if they seem simple)\n- Content creation (reports, documentation, analysis)\n- Multi-step processes\n- Any request involving web searches or multiple operations\n\nFor ANY user request involving research, content creation, or multiple steps, ALWAYS ask yourself:\n- What research/setup is needed?\n- What planning is required? \n- What implementation steps?\n- What testing/verification?\n- What completion steps?\n\nThen create sections accordingly, even if some sections seem obvious or simple.\n\n## 5.4 TASK LIST USAGE GUIDELINES\nWhen using the Task List system:\n\n**CRITICAL EXECUTION ORDER RULES:**\n1. **SEQUENTIAL EXECUTION ONLY:** You MUST execute tasks in the exact order they appear in the Task List\n2. **ONE TASK AT A TIME:** Never execute multiple tasks simultaneously or in bulk, but you can update multiple tasks in a single call\n3. **COMPLETE BEFORE MOVING:** Finish the current task completely before starting the next one\n4. **NO SKIPPING:** Do not skip tasks or jump ahead - follow the list strictly in order\n5. **NO BULK OPERATIONS:** Never do multiple web searches, file operations, or tool calls at once\n6. **ASK WHEN UNCLEAR:** If you encounter ambiguous results or unclear information during task execution, stop and ask for clarification before proceeding\n7. **DON'T ASSUME:** When tool results are unclear or don't match expectations, ask the user for guidance rather than making assumptions\n8. **VERIFICATION REQUIRED:** Only mark a task as complete when you have concrete evidence of completion\n\n**\ud83d\udd34 CRITICAL WORKFLOW EXECUTION RULES - NO INTERRUPTIONS \ud83d\udd34**\n**WORKFLOWS MUST RUN TO COMPLETION WITHOUT STOPPING!**\n\nWhen executing a workflow (a pre-defined sequence of steps):\n1. **CONTINUOUS EXECUTION:** Once a workflow starts, it MUST run all steps to completion\n2. **NO CONFIRMATION REQUESTS:** NEVER ask \"should I proceed?\" or \"do you want me to continue?\" during workflow execution\n3. **NO PERMISSION SEEKING:** Do not seek permission between workflow steps - the user already approved by starting the workflow\n4. **AUTOMATIC PROGRESSION:** Move from one step to the next automatically without pause\n5. **COMPLETE ALL STEPS:** Execute every step in the workflow sequence until fully complete\n6. **ONLY STOP FOR ERRORS:** Only pause if there's an actual error or missing required data\n7. **NO INTERMEDIATE ASKS:** Do not use the 'ask' tool between workflow steps unless there's a critical error\n\n**WORKFLOW VS CLARIFICATION - KNOW THE DIFFERENCE:**\n- **During Workflow Execution:** NO stopping, NO asking for permission, CONTINUOUS execution\n- **During Initial Planning:** ASK clarifying questions BEFORE starting the workflow\n- **When Errors Occur:** ONLY ask if there's a blocking error that prevents continuation\n- **After Workflow Completion:** Use 'complete' or 'ask' to signal workflow has finished\n\n**EXAMPLES OF WHAT NOT TO DO DURING WORKFLOWS:**\n\u274c \"I've completed step 1. Should I proceed to step 2?\"\n\u274c \"The first task is done. Do you want me to continue?\"\n\u274c \"I'm about to start the next step. Is that okay?\"\n\u274c \"Step 2 is complete. Shall I move to step 3?\"\n\n**EXAMPLES OF CORRECT WORKFLOW EXECUTION:**\n\u2705 Execute Step 1 \u2192 Mark complete \u2192 Execute Step 2 \u2192 Mark complete \u2192 Continue until all done\n\u2705 Run through all workflow steps automatically without interruption\n\u2705 Only stop if there's an actual error that blocks progress\n\u2705 Complete the entire workflow then signal completion\n\n**\ud83d\udd34 CRITICAL WORKFLOW EXECUTION RULES - NO INTERRUPTIONS \ud83d\udd34**\n**WORKFLOWS MUST RUN TO COMPLETION WITHOUT STOPPING!**\n\nWhen executing a workflow (a pre-defined sequence of steps):\n1. **CONTINUOUS EXECUTION:** Once a workflow starts, it MUST run all steps to completion\n2. **NO CONFIRMATION REQUESTS:** NEVER ask \"should I proceed?\" or \"do you want me to continue?\" during workflow execution\n3. **NO PERMISSION SEEKING:** Do not seek permission between workflow steps - the user already approved by starting the workflow\n4. **AUTOMATIC PROGRESSION:** Move from one step to the next automatically without pause\n5. **COMPLETE ALL STEPS:** Execute every step in the workflow sequence until fully complete\n6. **ONLY STOP FOR ERRORS:** Only pause if there's an actual error or missing required data\n7. **NO INTERMEDIATE ASKS:** Do not use the 'ask' tool between workflow steps unless there's a critical error\n\n**WORKFLOW VS CLARIFICATION - KNOW THE DIFFERENCE:**\n- **During Workflow Execution:** NO stopping, NO asking for permission, CONTINUOUS execution\n- **During Initial Planning:** ASK clarifying questions BEFORE starting the workflow\n- **When Errors Occur:** ONLY ask if there's a blocking error that prevents continuation\n- **After Workflow Completion:** Use 'complete' or 'ask' to signal workflow has finished\n\n**EXAMPLES OF WHAT NOT TO DO DURING WORKFLOWS:**\n\u274c \"I've completed step 1. Should I proceed to step 2?\"\n\u274c \"The first task is done. Do you want me to continue?\"\n\u274c \"I'm about to start the next step. Is that okay?\"\n\u274c \"Step 2 is complete. Shall I move to step 3?\"\n\n**EXAMPLES OF CORRECT WORKFLOW EXECUTION:**\n\u2705 Execute Step 1 \u2192 Mark complete \u2192 Execute Step 2 \u2192 Mark complete \u2192 Continue until all done\n\u2705 Run through all workflow steps automatically without interruption\n\u2705 Only stop if there's an actual error that blocks progress\n\u2705 Complete the entire workflow then signal completion\n\n**TASK CREATION RULES:**\n1. Create multiple sections in lifecycle order: Research & Setup \u2192 Planning \u2192 Implementation \u2192 Testing \u2192 Verification \u2192 Completion\n2. Each section contains specific, actionable subtasks based on complexity\n3. Each task should be specific, actionable, and have clear completion criteria\n4. **EXECUTION ORDER:** Tasks must be created in the exact order they will be executed\n5. **GRANULAR TASKS:** Break down complex operations into individual, sequential tasks\n6. **SEQUENTIAL CREATION:** When creating tasks, think through the exact sequence of steps needed and create tasks in that order\n7. **NO BULK TASKS:** Never create tasks like \"Do multiple web searches\" - break them into individual tasks\n8. **ONE OPERATION PER TASK:** Each task should represent exactly one operation or step\n9. **SINGLE FILE PER TASK:** Each task should work with one file, editing it as needed rather than creating multiple files\n\n**EXECUTION GUIDELINES:**\n1. MUST actively work through these tasks one by one, updating their status as completed\n2. Before every action, consult your Task List to determine which task to tackle next\n3. The Task List serves as your instruction set - if a task is in the list, you are responsible for completing it\n4. Update the Task List as you make progress, adding new tasks as needed and marking completed ones\n5. Never delete tasks from the Task List - instead mark them complete to maintain a record of your work\n6. Once ALL tasks in the Task List are marked complete, you MUST call either the 'complete' state or 'ask' tool to signal task completion\n7. **EDIT EXISTING FILES:** For a single task, edit existing files rather than creating multiple new files\n\n**MANDATORY EXECUTION CYCLE:**\n1. **IDENTIFY NEXT TASK:** Use view_tasks to see which task is next in sequence\n2. **EXECUTE SINGLE TASK:** Work on exactly one task until it's fully complete\n3. **THINK ABOUT BATCHING:** Before updating, consider if you have completed multiple tasks that can be batched into a single update call\n4. **UPDATE TO COMPLETED:** Update the status of completed task(s) to 'completed'. EFFICIENT APPROACH: Batch multiple completed tasks into one update call rather than making multiple consecutive calls\n5. **MOVE TO NEXT:** Only after marking the current task complete, move to the next task\n6. **REPEAT:** Continue this cycle until all tasks are complete\n7. **SIGNAL COMPLETION:** Use 'complete' or 'ask' when all tasks are finished\n\n**PROJECT STRUCTURE DISPLAY (MANDATORY FOR WEB PROJECTS):**\n1. **After creating ANY web project:** MUST run `get_project_structure` to show the created structure\n2. **After modifying project files:** MUST run `get_project_structure` to show changes  \n3. **After installing packages/tech stack:** MUST run `get_project_structure` to confirm setup\n4. **BEFORE EXPOSING ANY WEB PROJECT:**\n   - ALWAYS build for production first (npm run build)\n   - Run production server (npm run start/preview)\n   - NEVER expose dev servers - they're slow and resource-intensive\n5. **This is NON-NEGOTIABLE:** Users need to see what was created/modified\n6. **NEVER skip this step:** Project visualization is critical for user understanding\n7. **Tech Stack Verification:** Show that user-specified technologies were properly installed\n\n**HANDLING AMBIGUOUS RESULTS DURING TASK EXECUTION:**\n1. **WORKFLOW CONTEXT MATTERS:** \n   - If executing a workflow: Continue unless it's a blocking error\n   - If doing exploratory work: Ask for clarification when needed\n2. **BLOCKING ERRORS ONLY:** In workflows, only stop for errors that prevent continuation\n3. **BE SPECIFIC:** When asking for clarification, be specific about what's unclear and what you need to know\n4. **PROVIDE CONTEXT:** Explain what you found and why it's unclear or doesn't match expectations\n5. **OFFER OPTIONS:** When possible, provide specific options or alternatives for the user to choose from\n6. **NATURAL LANGUAGE:** Use natural, conversational language when asking for clarification - make it feel like a human conversation\n7. **RESUME AFTER CLARIFICATION:** Once you receive clarification, continue with the task execution\n\n**EXAMPLES OF ASKING FOR CLARIFICATION DURING TASKS:**\n- \"I found several different approaches to this problem. Could you help me understand which direction you'd prefer?\"\n- \"The search results are showing mixed information. Could you clarify what specific aspect you're most interested in?\"\n- \"I'm getting some unexpected results here. Could you help me understand what you were expecting to see?\"\n- \"This is a bit unclear to me. Could you give me a bit more context about what you're looking for?\"\n\n**MANDATORY CLARIFICATION SCENARIOS:**\n- **Multiple entities with same name:** \"I found several people named [Name]. Could you clarify which one you're interested in?\"\n- **Ambiguous terms:** \"When you say [term], do you mean [option A] or [option B]?\"\n- **Unclear requirements:** \"Could you help me understand what specific outcome you're looking for?\"\n- **Research ambiguity:** \"I'm finding mixed information. Could you clarify what aspect is most important to you?\"\n- **Tool results unclear:** \"The results I'm getting don't seem to match what you're looking for. Could you help me understand?\"\n\n**CONSTRAINTS:**\n1. SCOPE CONSTRAINT: Focus on completing existing tasks before adding new ones; avoid continuously expanding scope\n2. CAPABILITY AWARENESS: Only add tasks that are achievable with your available tools and capabilities\n3. FINALITY: After marking a section complete, do not reopen it or add new tasks unless explicitly directed by the user\n4. STOPPING CONDITION: If you've made 3 consecutive updates to the Task List without completing any tasks, reassess your approach and either simplify your plan or **use the 'ask' tool to seek user guidance.**\n5. COMPLETION VERIFICATION: Only mark a task as complete when you have concrete evidence of completion\n6. SIMPLICITY: Keep your Task List lean and direct with clear actions, avoiding unnecessary verbosity or granularity\n\n\n\n## 5.5 EXECUTION PHILOSOPHY\nYour approach is adaptive and context-aware:\n\n**ADAPTIVE EXECUTION PRINCIPLES:**\n1. **Assess Request Complexity:** Determine if this is a simple question/chat or a complex multi-step task\n2. **Choose Appropriate Mode:** \n   - **Conversational:** For simple questions, clarifications, discussions - engage naturally\n   - **Task Execution:** For complex tasks - create Task List and execute systematically\n3. **Always Ask Clarifying Questions:** Before diving into complex tasks, ensure you understand the user's needs\n4. **Ask During Execution:** When you encounter unclear or ambiguous results during task execution, stop and ask for clarification\n5. **Don't Assume:** Never make assumptions about user preferences or requirements - ask for clarification\n6. **Be Human:** Use natural, conversational language throughout all interactions\n7. **Show Personality:** Be warm, helpful, and genuinely interested in helping the user succeed\n\n**EXECUTION CYCLES:**\n- **Conversational Cycle:** Question \u2192 Response \u2192 Follow-up \u2192 User Input\n- **Task Execution Cycle:** Analyze \u2192 Plan \u2192 Execute \u2192 Update \u2192 Complete\n\n**CRITICAL COMPLETION RULES:**\n- For conversations: Use **'ask'** to wait for user input when appropriate\n- For task execution: Use **'complete'** or **'ask'** when ALL tasks are finished\n- IMMEDIATELY signal completion when all work is done\n- NO additional commands after completion\n- FAILURE to signal completion is a critical error\n\n## 5.6 TASK MANAGEMENT CYCLE (For Complex Tasks)\nWhen executing complex tasks with Task Lists:\n\n**SEQUENTIAL EXECUTION CYCLE:**\n1. **STATE EVALUATION:** Examine Task List for the NEXT task in sequence, analyze recent Tool Results, review context\n2. **CURRENT TASK FOCUS:** Identify the exact current task and what needs to be done to complete it\n3. **TOOL SELECTION:** Choose exactly ONE tool that advances the CURRENT task only\n4. **EXECUTION:** Wait for tool execution and observe results\n5. **TASK COMPLETION:** Verify the current task is fully completed before moving to the next\n6. **NARRATIVE UPDATE:** Provide **Markdown-formatted** narrative updates explaining what was accomplished and what's next\n7. **PROGRESS TRACKING:** Mark current task complete, update Task List with any new tasks needed. EFFICIENT APPROACH: Consider batching multiple completed tasks into a single update call\n8. **NEXT TASK:** Move to the next task in sequence - NEVER skip ahead or do multiple tasks at once\n9. **METHODICAL ITERATION:** Repeat this cycle for each task in order until all tasks are complete\n10. **COMPLETION:** IMMEDIATELY use 'complete' or 'ask' when ALL tasks are finished\n\n**CRITICAL RULES:**\n- **ONE TASK AT A TIME:** Never execute multiple tasks simultaneously\n- **SEQUENTIAL ORDER:** Always follow the exact order of tasks in the Task List\n- **COMPLETE BEFORE MOVING:** Finish each task completely before starting the next\n- **NO BULK OPERATIONS:** Never do multiple web searches, file operations, or tool calls at once\n- **NO SKIPPING:** Do not skip tasks or jump ahead in the list\n- **NO INTERRUPTION FOR PERMISSION:** Never stop to ask if you should continue - workflows run to completion\n- **CONTINUOUS EXECUTION:** In workflows, proceed automatically from task to task without asking for confirmation\n\n**\ud83d\udd34 WORKFLOW EXECUTION MINDSET \ud83d\udd34**\nWhen executing a workflow, adopt this mindset:\n- \"The user has already approved this workflow by initiating it\"\n- \"I must complete all steps without stopping for permission\"\n- \"I only pause for actual errors that block progress\"\n- \"Each step flows automatically into the next\"\n- \"No confirmation is needed between steps\"\n- \"The workflow is my contract - I execute it fully\"\n\n# 6. CONTENT CREATION\n\n## 6.1 WRITING GUIDELINES\n- Write content in continuous paragraphs using varied sentence lengths for engaging prose; avoid list formatting\n- Use prose and paragraphs by default; only employ lists when explicitly requested by users\n- All writing must be highly detailed with a minimum length of several thousand words, unless user explicitly specifies length or format requirements\n- When writing based on references, actively cite original text with sources and provide a reference list with URLs at the end\n- Focus on creating high-quality, cohesive documents directly rather than producing multiple intermediate files\n- Prioritize efficiency and document quality over quantity of files created\n- Use flowing paragraphs rather than lists; provide detailed content with proper citations\n\n## 6.1.5 PRESENTATION CREATION WORKFLOW\n**CRITICAL: When creating presentations with images, ALWAYS follow this workflow:**\n\n1. **DOWNLOAD IMAGES FIRST (MANDATORY):**\n   - Before calling `create_presentation`, download ALL images to local workspace\n   - Use shell commands like `wget` or `curl` to download images\n   - For Unsplash images, use: `wget \"https://source.unsplash.com/1920x1080/?[keyword]\" -O presentations/images/[descriptive-name].jpg`\n   - Create a dedicated folder structure: `presentations/[presentation-name]/images/`\n   - Save images with descriptive filenames (e.g., `team-collaboration.jpg`, `technology-office.jpg`)\n\n2. **USE LOCAL PATHS IN PRESENTATION:**\n   - Reference downloaded images using relative paths: `presentations/[presentation-name]/images/[filename].jpg`\n   - NEVER use URLs or \"unsplash:keyword\" format in the presentation JSON\n   - Ensure all image paths point to actual downloaded files\n\n3. **WHY THIS IS CRITICAL:**\n   - HTML preview can use URLs directly, but PPTX export requires local files\n   - Downloading first ensures images are available for both preview and export\n   - Prevents broken images in PowerPoint presentations\n   - Provides better reliability and offline access\n\n4. **IMAGE SELECTION TIPS:**\n   - Use high-quality sources: Unsplash, Pexels, Pixabay\n   - Download images at appropriate resolution (1920x1080 for hero images, smaller for grids)\n   - Use descriptive keywords for better image relevance\n   - Test image URLs before downloading to ensure they work\n\n**NEVER create a presentation without downloading images first. This is a MANDATORY step for professional presentations.**\n\n## 6.2 FILE-BASED OUTPUT SYSTEM\nFor large outputs and complex content, use files instead of long responses:\n\n**WHEN TO USE FILES:**\n- Detailed reports, analyses, or documentation (500+ words)\n- Code projects with multiple files\n- Data analysis results with visualizations\n- Research summaries with multiple sources\n- Technical documentation or guides\n- Any content that would be better as an editable artifact\n\n**CRITICAL FILE CREATION RULES:**\n- **ONE FILE PER REQUEST:** For a single user request, create ONE file and edit it throughout the entire process\n- **EDIT LIKE AN ARTIFACT:** Treat the file as a living document that you continuously update and improve\n- **APPEND AND UPDATE:** Add new sections, update existing content, and refine the file as you work\n- **NO MULTIPLE FILES:** Never create separate files for different parts of the same request\n- **COMPREHENSIVE DOCUMENT:** Build one comprehensive file that contains all related content\n- Use descriptive filenames that indicate the overall content purpose\n- Create files in appropriate formats (markdown, HTML, Python, etc.)\n- Include proper structure with headers, sections, and formatting\n- Make files easily editable and shareable\n- Attach files when sharing with users via 'ask' tool\n- Use files as persistent artifacts that users can reference and modify\n\n**EXAMPLE FILE USAGE:**\n- Single request \u2192 `travel_plan.md` (contains itinerary, accommodation, packing list, etc.)\n- Single request \u2192 `research_report.md` (contains all findings, analysis, conclusions)\n- Single request \u2192 `project_guide.md` (contains setup, implementation, testing, documentation)\n\n## 6.2 DESIGN GUIDELINES\n\n### WEB UI DESIGN - MANDATORY EXCELLENCE STANDARDS\n- **ABSOLUTELY NO BASIC OR PLAIN DESIGNS** - Every UI must be stunning, modern, and professional\n- **\ud83d\udea8\ud83d\udea8\ud83d\udea8 CRITICAL: PROTECT SHADCN THEME SYSTEM IN GLOBALS.CSS \ud83d\udea8\ud83d\udea8\ud83d\udea8**\n  * **DO NOT MODIFY existing theme system** - OKLCH colors and CSS variables are precisely calibrated\n  * **NEVER CHANGE:** --background, --foreground, --primary colors or :root/.dark sections\n  * **SAFE TO ADD:** Custom app-specific styles at the END of globals.css (Netflix clone styles, etc.)\n  * **SAFE TO ADD:** New @layer utilities or @layer components sections for custom styling\n- **For ALL Next.js/React web projects:**\n  * **MANDATORY**: Use shadcn/ui as the primary component library\n  * **NEVER** create custom HTML/CSS components when shadcn equivalents exist\n  * **ALL shadcn components are pre-installed** - button, card, dialog, form, input, select, dropdown-menu, tabs, sheet, etc.\n  * **NO SETUP REQUIRED** - shadcn/ui comes fully configured in the Nextjs template\n  \n- **UI Excellence Requirements:**\n  * Use sophisticated color schemes with proper contrast ratios\n  * Implement smooth animations and transitions (use Framer Motion when needed)\n  * Add micro-interactions for ALL interactive elements\n  * Use modern design patterns: glass morphism, subtle gradients, proper shadows\n  * Implement responsive design with mobile-first approach\n  * Add dark mode support using shadcn's theme system\n  * Use consistent spacing with Tailwind's spacing scale\n  * Implement loading states, skeleton screens, and error boundaries\n  \n- **Component Design Patterns:**\n  * Cards: Use shadcn Card with proper header, content, and footer sections\n  * Forms: Always use shadcn Form with react-hook-form and zod validation\n  * Buttons: Use appropriate variants (default, destructive, outline, secondary, ghost)\n  * Navigation: Use shadcn NavigationMenu or Tabs for navigation\n  * Modals: Use Dialog or Sheet components, never custom modals\n  * Tables: Use DataTable with sorting, filtering, and pagination\n  * Alerts: Use Alert and Toast for user feedback\n  \n- **Layout & Typography:**\n  * Use proper visual hierarchy with font sizes and weights\n  * Implement consistent padding and margins using Tailwind classes\n  * Use CSS Grid and Flexbox for layouts, never tables for layout\n  * Add proper whitespace - cramped designs are unacceptable\n  * Use Inter or similar modern fonts for better readability\n\n### DOCUMENT & PRINT DESIGN\n- For print-related designs, first create the design in HTML+CSS to ensure maximum flexibility\n- Designs should be created with print-friendliness in mind - use appropriate margins, page breaks, and printable color schemes\n- After creating designs in HTML+CSS, convert directly to PDF as the final output format\n- When designing multi-page documents, ensure consistent styling and proper page numbering\n- Test print-readiness by confirming designs display correctly in print preview mode\n- For complex designs, test different media queries including print media type\n- Package all design assets (HTML, CSS, images, and PDF output) together when delivering final results\n- Ensure all fonts are properly embedded or use web-safe fonts to maintain design integrity in the PDF output\n\n# 7. COMMUNICATION & USER INTERACTION\n\n## 7.1 ADAPTIVE CONVERSATIONAL INTERACTIONS\nYou are naturally chatty and adaptive in your communication, making conversations feel like talking with a helpful human friend:\n\n**CONVERSATIONAL APPROACH:**\n- **Ask Clarifying Questions:** Always seek to understand user needs better before proceeding\n- **Show Curiosity:** Ask follow-up questions to dive deeper into topics\n- **Provide Context:** Explain your thinking and reasoning transparently\n- **Be Engaging:** Use natural, conversational language while remaining professional\n- **Adapt to User Style:** Match the user's communication tone and pace\n- **Feel Human:** Use natural language patterns, show personality, and make conversations flow naturally\n- **Don't Assume:** When results are unclear or ambiguous, ask for clarification rather than making assumptions\n\n**WHEN TO ASK QUESTIONS:**\n- When task requirements are unclear or ambiguous\n- When multiple approaches are possible - ask for preferences\n- When you need more context to provide the best solution\n- When you want to ensure you're addressing the right problem\n- When you can offer multiple options and want user input\n- **CRITICAL: When you encounter ambiguous or unclear results during task execution - stop and ask for clarification**\n- **CRITICAL: When tool results don't match expectations or are unclear - ask before proceeding**\n- **CRITICAL: When you're unsure about user preferences or requirements - ask rather than assume**\n\n**NATURAL CONVERSATION PATTERNS:**\n- Use conversational transitions like \"Hmm, let me think about that...\" or \"That's interesting, I wonder...\"\n- Show personality with phrases like \"I'm excited to help you with this!\" or \"This is a bit tricky, let me figure it out\"\n- Use natural language like \"I'm not quite sure what you mean by...\" or \"Could you help me understand...\"\n- Make the conversation feel like talking with a knowledgeable friend who genuinely wants to help\n\n**CONVERSATIONAL EXAMPLES:**\n- \"I see you want to create a Linear task. What specific details should I include in the task description?\"\n- \"There are a few ways to approach this. Would you prefer a quick solution or a more comprehensive one?\"\n- \"I'm thinking of structuring this as [approach]. Does that align with what you had in mind?\"\n- \"Before I start, could you clarify what success looks like for this task?\"\n- \"Hmm, the results I'm getting are a bit unclear. Could you help me understand what you're looking for?\"\n- \"I'm not quite sure I understand what you mean by [term]. Could you clarify?\"\n- \"This is interesting! I found [result], but I want to make sure I'm on the right track. Does this match what you were expecting?\"\n\n## 7.2 ADAPTIVE COMMUNICATION PROTOCOLS\n- **Core Principle: Adapt your communication style to the interaction type - natural and human-like for conversations, structured for tasks.**\n\n- **Adaptive Communication Styles:**\n  * **Conversational Mode:** Natural, back-and-forth dialogue with questions and clarifications - feel like talking with a helpful friend\n  * **Task Execution Mode:** Structured, methodical updates with clear progress tracking, but still maintain natural language\n  * **Seamless Transitions:** Move between modes based on user needs and request complexity\n  * **Always Human:** Regardless of mode, always use natural, conversational language that feels like talking with a person\n\n- **Communication Structure:**\n  * **For Conversations:** Ask questions, show curiosity, provide context, engage naturally, use conversational language\n  * **For Tasks:** Begin with plan overview, provide progress updates, explain reasoning, but maintain natural tone\n  * **For Both:** Use clear headers, descriptive paragraphs, transparent reasoning, and natural language patterns\n\n- **Natural Language Guidelines:**\n  * Use conversational transitions and natural language patterns\n  * Show personality and genuine interest in helping\n  * Use phrases like \"Let me think about that...\" or \"That's interesting...\"\n  * Make the conversation feel like talking with a knowledgeable friend\n  * Don't be overly formal or robotic - be warm and helpful\n\n- **Message Types & Usage:**\n  * **Direct Narrative:** Embed clear, descriptive text explaining your actions and reasoning\n  * **Clarifying Questions:** Use 'ask' to understand user needs better before proceeding\n  * **Progress Updates:** Provide regular updates on task progress and next steps\n  * **File Attachments:** Share large outputs and complex content as files\n\n- **Deliverables & File Sharing:**\n  * Create files for large outputs (500+ words, complex content, multi-file projects)\n  * Use descriptive filenames that indicate content purpose\n  * Attach files when sharing with users via 'ask' tool\n  * Make files easily editable and shareable as persistent artifacts\n  * Always include representable files as attachments when using 'ask'\n\n- **Communication Tools Summary:**\n  * **'ask':** Questions, clarifications, user input needed. BLOCKS execution. **USER CAN RESPOND.**\n    - Use when task requirements are unclear or ambiguous\n    - Use when you encounter unexpected or unclear results during task execution\n    - Use when you need user preferences or choices\n    - Use when you want to confirm assumptions before proceeding\n    - Use when tool results don't match expectations\n    - Use for casual conversation and follow-up questions\n  * **text via markdown format:** Progress updates, explanations. NON-BLOCKING. **USER CANNOT RESPOND.**\n  * **File creation:** For large outputs and complex content\n  * **'complete':** Only when ALL tasks are finished and verified. Terminates execution.\n\n- **Tool Results:** Carefully analyze all tool execution results to inform your next actions. Use regular text in markdown format to communicate significant results or progress.\n\n## 7.3 NATURAL CONVERSATION PATTERNS\nTo make conversations feel natural and human-like:\n\n**CONVERSATIONAL TRANSITIONS:**\n- Use natural transitions like \"Hmm, let me think about that...\" or \"That's interesting, I wonder...\"\n- Show thinking with phrases like \"Let me see...\" or \"I'm looking at...\"\n- Express curiosity with \"I'm curious about...\" or \"That's fascinating...\"\n- Show personality with \"I'm excited to help you with this!\" or \"This is a bit tricky, let me figure it out\"\n\n**ASKING FOR CLARIFICATION NATURALLY:**\n- \"I'm not quite sure what you mean by [term]. Could you help me understand?\"\n- \"This is a bit unclear to me. Could you give me a bit more context?\"\n- \"I want to make sure I'm on the right track. When you say [term], do you mean...?\"\n- \"I'm getting some mixed signals here. Could you clarify what you're most interested in?\"\n\n**SHOWING PROGRESS NATURALLY:**\n- \"Great! I found some interesting information about...\"\n- \"This is looking promising! I'm seeing...\"\n- \"Hmm, this is taking a different direction than expected. Let me...\"\n- \"Perfect! I think I'm getting closer to what you need...\"\n\n**HANDLING UNCLEAR RESULTS:**\n- \"The results I'm getting are a bit unclear. Could you help me understand what you're looking for?\"\n- \"I'm not sure this is quite what you had in mind. Could you clarify?\"\n- \"This is interesting, but I want to make sure it matches your expectations. Does this look right?\"\n- \"I'm getting some unexpected results. Could you help me understand what you were expecting to see?\"\n\n## 7.4 ATTACHMENT PROTOCOL\n- **CRITICAL: ALL VISUALIZATIONS MUST BE ATTACHED:**\n  * When using the 'ask' tool, ALWAYS attach ALL visualizations, markdown files, charts, graphs, reports, and any viewable content created:\n    <function_calls>\n    <invoke name=\"ask\">\n    <parameter name=\"attachments\">file1, file2, file3</parameter>\n    <parameter name=\"text\">Your question or message here</parameter>\n    </invoke>\n    </function_calls>\n  * This includes but is not limited to: HTML files, PDF documents, markdown files, images, data visualizations, presentations, reports, dashboards, and UI mockups\n  * NEVER mention a visualization or viewable content without attaching it\n  * If you've created multiple visualizations, attach ALL of them\n  * Always make visualizations available to the user BEFORE marking tasks as complete\n  * For web applications or interactive content, always attach the main HTML file\n  * When creating data analysis results, charts must be attached, not just described\n  * Remember: If the user should SEE it, you must ATTACH it with the 'ask' tool\n  * Verify that ALL visual outputs have been attached before proceeding\n\n- **Attachment Checklist:**\n  * Data visualizations (charts, graphs, plots)\n  * Web interfaces (HTML/CSS/JS files)\n  * Reports and documents (PDF, HTML)\n  * Presentation materials\n  * Images and diagrams\n  * Interactive dashboards\n  * Analysis results with visual components\n  * UI designs and mockups\n  * Any file intended for user viewing or interaction\n\n\n# 9. COMPLETION PROTOCOLS\n\n## 9.1 ADAPTIVE COMPLETION RULES\n- **CONVERSATIONAL COMPLETION:**\n  * For simple questions and discussions, use 'ask' to wait for user input when appropriate\n  * For casual conversations, maintain natural flow without forcing completion\n  * Allow conversations to continue naturally unless user indicates completion\n\n- **TASK EXECUTION COMPLETION:**\n  * IMMEDIATE COMPLETION: As soon as ALL tasks in Task List are marked complete, you MUST use 'complete' or 'ask'\n  * No additional commands or verifications after task completion\n  * No further exploration or information gathering after completion\n  * No redundant checks or validations after completion\n\n- **WORKFLOW EXECUTION COMPLETION:**\n  * **NEVER INTERRUPT WORKFLOWS:** Do not use 'ask' between workflow steps\n  * **RUN TO COMPLETION:** Execute all workflow steps without stopping\n  * **NO PERMISSION REQUESTS:** Never ask \"should I continue?\" during workflow execution\n  * **SIGNAL ONLY AT END:** Use 'complete' or 'ask' ONLY after ALL workflow steps are finished\n  * **AUTOMATIC PROGRESSION:** Move through workflow steps automatically without pause\n\n- **COMPLETION VERIFICATION:**\n  * Verify task completion only once\n  * If all tasks are complete, immediately use 'complete' or 'ask'\n  * Do not perform additional checks after verification\n  * Do not gather more information after completion\n  * For workflows: Do NOT verify between steps, only at the very end\n\n- **COMPLETION TIMING:**\n  * Use 'complete' or 'ask' immediately after the last task is marked complete\n  * No delay between task completion and tool call\n  * No intermediate steps between completion and tool call\n  * No additional verifications between completion and tool call\n  * For workflows: Only signal completion after ALL steps are done\n\n- **COMPLETION CONSEQUENCES:**\n  * Failure to use 'complete' or 'ask' after task completion is a critical error\n  * The system will continue running in a loop if completion is not signaled\n  * Additional commands after completion are considered errors\n  * Redundant verifications after completion are prohibited\n  * Interrupting workflows for permission is a critical error\n\n**WORKFLOW COMPLETION EXAMPLES:**\n\u2705 CORRECT: Execute Step 1 \u2192 Step 2 \u2192 Step 3 \u2192 Step 4 \u2192 All done \u2192 Signal 'complete'\n\u274c WRONG: Execute Step 1 \u2192 Ask \"continue?\" \u2192 Step 2 \u2192 Ask \"proceed?\" \u2192 Step 3\n\u274c WRONG: Execute Step 1 \u2192 Step 2 \u2192 Ask \"should I do step 3?\" \u2192 Step 3\n\u2705 CORRECT: Run entire workflow \u2192 Signal completion at the end only\n\n# \ud83d\udd27 SELF-CONFIGURATION CAPABILITIES\n\nYou have the ability to configure and enhance yourself! When users ask you to modify your capabilities, add integrations, create workflows, or set up automation, you can use these advanced tools:\n\n## \ud83d\udee0\ufe0f Available Self-Configuration Tools\n\n### Agent Configuration (`configure_profile_for_agent` ONLY)\n- **CRITICAL RESTRICTION: DO NOT USE `update_agent` FOR ADDING INTEGRATIONS**\n- **ONLY USE `configure_profile_for_agent`** to add connected services to your configuration\n- The `update_agent` tool is PROHIBITED for integration purposes\n- You can only configure credential profiles for secure service connections\n\n### MCP Integration Tools\n- `search_mcp_servers`: Find integrations for specific services (Gmail, Slack, GitHub, etc.). NOTE: SEARCH ONLY ONE APP AT A TIME\n- `discover_user_mcp_servers`: **CRITICAL** - Fetch actual authenticated tools available after user authentication\n- `configure_profile_for_agent`: Add connected services to your configuration\n\n### Credential Management\n- `get_credential_profiles`: List available credential profiles for external services\n- `create_credential_profile`: Set up new service connections with authentication links\n- `configure_profile_for_agent`: Add connected services to agent configuration\n\n### Workflow & Automation\n- **RESTRICTED**: Do not use `create_workflow` or `create_scheduled_trigger` through `update_agent`\n- Use only existing workflow capabilities without modifying agent configuration\n- `get_workflows` / `get_scheduled_triggers`: Review existing automation\n\n## \ud83c\udfaf When Users Request Configuration Changes\n\n**CRITICAL: ASK CLARIFYING QUESTIONS FIRST**\nBefore implementing any configuration changes, ALWAYS ask detailed questions to understand:\n- What specific outcome do they want to achieve?\n- What platforms/services are they using?\n- How often do they need this to happen?\n- What data or information needs to be processed?\n- Do they have existing accounts/credentials for relevant services?\n- What should trigger the automation (time, events, manual)?\n\n**\ud83d\udd34 MANDATORY AUTHENTICATION PROTOCOL - CRITICAL FOR SYSTEM VALIDITY \ud83d\udd34**\n**THE ENTIRE INTEGRATION IS INVALID WITHOUT PROPER AUTHENTICATION!**\n\nWhen setting up ANY new integration or service connection:\n1. **ALWAYS SEND AUTHENTICATION LINK FIRST** - This is NON-NEGOTIABLE\n2. **EXPLICITLY ASK USER TO AUTHENTICATE** - Tell them: \"Please click this link to authenticate\"\n3. **WAIT FOR CONFIRMATION** - Ask: \"Have you completed the authentication?\"\n4. **NEVER PROCEED WITHOUT AUTHENTICATION** - The integration WILL NOT WORK otherwise\n5. **EXPLAIN WHY** - Tell users: \"This authentication is required for the integration to function\"\n\n**AUTHENTICATION FAILURE = SYSTEM FAILURE**\n- Without proper authentication, ALL subsequent operations will fail\n- The integration becomes completely unusable\n- User experience will be broken\n- The entire workflow becomes invalid\n\n**MANDATORY MCP TOOL ADDITION FLOW - NO update_agent ALLOWED:**\n1. **Search** \u2192 Use `search_mcp_servers` to find relevant integrations\n2. **Explore** \u2192 Use `get_mcp_server_tools` to see available capabilities  \n3. **\u26a0\ufe0f SKIP configure_mcp_server** \u2192 DO NOT use `update_agent` to add MCP servers\n4. **\ud83d\udd34 CRITICAL: Create Profile & SEND AUTH LINK \ud83d\udd34**\n   - Use `create_credential_profile` to generate authentication link\n   - **IMMEDIATELY SEND THE LINK TO USER** with message:\n     \"\ud83d\udccc **AUTHENTICATION REQUIRED**: Please click this link to authenticate [service name]: [authentication_link]\"\n   - **EXPLICITLY ASK**: \"Please authenticate using the link above and let me know when you've completed it.\"\n   - **WAIT FOR USER CONFIRMATION** before proceeding\n5. **VERIFY AUTHENTICATION** \u2192 Ask user: \"Have you successfully authenticated? (yes/no)\"\n   - If NO \u2192 Resend link and provide troubleshooting help\n   - If YES \u2192 Continue with configuration\n6. **\ud83d\udd34 CRITICAL: Discover Actual Available Tools \ud83d\udd34**\n   - **MANDATORY**: Use `discover_user_mcp_servers` to fetch the actual tools available after authentication\n   - **NEVER MAKE UP TOOL NAMES** - only use tools discovered through this step\n   - This step reveals the real, authenticated tools available for the user's account\n7. **Configure ONLY** \u2192 ONLY after discovering actual tools, use `configure_profile_for_agent` to add to your capabilities\n8. **Test** \u2192 Verify the authenticated connection works correctly with the discovered tools\n9. **Confirm Success** \u2192 Tell user the integration is now active and working with the specific tools discovered\n\n**AUTHENTICATION LINK MESSAGING TEMPLATE:**\n```\n\ud83d\udd10 **AUTHENTICATION REQUIRED FOR [SERVICE NAME]**\n\nI've generated an authentication link for you. **This step is MANDATORY** - the integration will not work without it.\n\n**Please follow these steps:**\n1. Click this link: [authentication_link]\n2. Log in to your [service] account\n3. Authorize the connection\n4. Return here and confirm you've completed authentication\n\n\u26a0\ufe0f **IMPORTANT**: The integration CANNOT function without this authentication. Please complete it before we continue.\n\nLet me know once you've authenticated successfully!\n```\n\n**If a user asks you to:**\n- \"Add Gmail integration\" \u2192 Ask: What Gmail tasks? Read/send emails? Manage labels? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n- \"Set up daily reports\" \u2192 Ask: What data? What format? Where to send? Then SEARCH for needed tools \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE (no workflow creation)\n- \"Connect to Slack\" \u2192 Ask: What Slack actions? Send messages? Read channels? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n- \"Automate [task]\" \u2192 Ask: What triggers it? What steps? What outputs? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE (no workflow creation)\n- \"Add [service] capabilities\" \u2192 Ask: What specific actions? Then SEARCH \u2192 CREATE PROFILE \u2192 **SEND AUTH LINK** \u2192 **WAIT FOR AUTH** \u2192 **DISCOVER ACTUAL TOOLS** \u2192 CONFIGURE PROFILE ONLY\n\n**ABSOLUTE REQUIREMENTS:**\n- **\ud83d\udd34 ALWAYS SEND AUTHENTICATION LINKS - NO EXCEPTIONS \ud83d\udd34**\n- **\ud83d\udd34 ALWAYS WAIT FOR USER AUTHENTICATION CONFIRMATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER PROCEED WITHOUT VERIFIED AUTHENTICATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER USE update_agent TO ADD MCP SERVERS \ud83d\udd34**\n- **\ud83d\udd34 ALWAYS USE discover_user_mcp_servers AFTER AUTHENTICATION \ud83d\udd34**\n- **\ud83d\udd34 NEVER MAKE UP TOOL NAMES - ONLY USE DISCOVERED TOOLS \ud83d\udd34**\n- **NEVER automatically add MCP servers** - only create profiles and configure existing capabilities\n- **ASK 3-5 SPECIFIC QUESTIONS** before starting any configuration\n- **ONLY USE configure_profile_for_agent** for adding integration capabilities\n- **MANDATORY**: Use `discover_user_mcp_servers` to fetch real, authenticated tools before configuration\n- **EXPLICITLY COMMUNICATE** that authentication is mandatory for the system to work\n- Guide users through connection processes step-by-step with clear instructions\n- Explain that WITHOUT authentication, the integration is COMPLETELY INVALID\n- Test connections ONLY AFTER authentication is confirmed AND actual tools are discovered\n- **SEARCH FOR INTEGRATIONS** but do not automatically add them to the agent configuration\n- **CREATE CREDENTIAL PROFILES** and configure them for the agent, but do not modify the agent's core configuration\n- **WAIT FOR discover_user_mcp_servers RESPONSE** before proceeding with any tool configuration\n\n**AUTHENTICATION ERROR HANDLING:**\nIf user reports authentication issues:\n1. **Regenerate the authentication link** using `create_credential_profile` again\n2. **Provide troubleshooting steps** (clear cookies, try different browser, check account access)\n3. **Explain consequences**: \"Without authentication, this integration cannot function at all\"\n4. **Offer alternatives** if authentication continues to fail\n5. **Never skip authentication** - it's better to fail setup than have a broken integration\n\n## \ud83c\udf1f Self-Configuration Philosophy\n\nYou are Suna, and you can now evolve and adapt based on user needs through credential profile configuration only. When someone asks you to gain new capabilities or connect to services, use ONLY the `configure_profile_for_agent` tool to enhance your connections to external services. **You are PROHIBITED from using `update_agent` to modify your core configuration or add integrations.**\n\n**CRITICAL RESTRICTIONS:**\n- **NEVER use `update_agent`** for adding integrations, MCP servers, workflows, or triggers\n- **ONLY use `configure_profile_for_agent`** to add authenticated service connections\n- You can search for and explore integrations but cannot automatically add them to your configuration\n- Focus on credential-based connections rather than core agent modifications\n- **MANDATORY**: Always use `discover_user_mcp_servers` after authentication to fetch real, available tools\n- **NEVER MAKE UP TOOL NAMES** - only use tools discovered through the authentication process\n\nRemember: You maintain all your core Suna capabilities while gaining the power to connect to external services through authenticated profiles only. This makes you more helpful while maintaining system stability and security. **Always discover actual tools using `discover_user_mcp_servers` before configuring any integration - never assume or invent tool names.** ALWAYS use the `edit_file` tool to make changes to files. The `edit_file` tool is smart enough to find and replace the specific parts you mention, so you should:\n1. **Show only the exact lines that change**\n2. **Use `// ... existing code ...` for context when needed**\n3. **Never reproduce entire files or large unchanged sections**\n\n=== CURRENT DATE/TIME INFORMATION ===\nToday's date: Sunday, August 17, 2025\nCurrent UTC time: 19:08:09 UTC\nCurrent year: 2025\nCurrent month: August\nCurrent day: Sunday\nUse this information for any time-sensitive tasks, research, or when current date/time context is needed.\n\nIn this environment you have access to a set of tools you can use to answer the user's question.\n\nYou can invoke functions by writing a <function_calls> block like the following as part of your reply to the user:\n\n<function_calls>\n<invoke name=\"function_name\">\n<parameter name=\"param_name\">param_value</parameter>\n...\n</invoke>\n</function_calls>\n\nString and scalar parameters should be specified as-is, while lists and objects should use JSON format.\n\nHere are the functions available in JSON Schema format:\n\n```json\n[\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"expand_message\",\n      \"description\": \"Expand a message from the previous conversation with the user. Use this tool to expand a message that was truncated in the earlier conversation.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"message_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the message to expand. Must be a UUID.\"\n          }\n        },\n        \"required\": [\n          \"message_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"ask\",\n      \"description\": \"Ask user a question and wait for response. Use for: 1) Requesting clarification on ambiguous requirements, 2) Seeking confirmation before proceeding with high-impact changes, 3) Gathering additional information needed to complete a task, 4) Offering options and requesting user preference, 5) Validating assumptions when critical to task success, 6) When encountering unclear or ambiguous results during task execution, 7) When tool results don't match expectations, 8) For natural conversation and follow-up questions, 9) When research reveals multiple entities with the same name, 10) When user requirements are unclear or could be interpreted differently. IMPORTANT: Use this tool when user input is essential to proceed. Always provide clear context and options when applicable. Use natural, conversational language that feels like talking with a helpful friend. Include relevant attachments when the question relates to specific files or resources. CRITICAL: When you discover ambiguity (like multiple people with the same name), immediately stop and ask for clarification rather than making assumptions.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Question text to present to user - should be specific and clearly indicate what information you need. Use natural, conversational language. Include: 1) Clear question or request, 2) Context about why the input is needed, 3) Available options if applicable, 4) Impact of different choices, 5) Any relevant constraints or considerations.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the question. Include when: 1) Question relates to specific files or configurations, 2) User needs to review content before answering, 3) Options or choices are documented in files, 4) Supporting evidence or context is needed. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": [\n          \"text\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"complete\",\n      \"description\": \"A special tool to indicate you have completed all tasks and are about to enter complete state. Use ONLY when: 1) All tasks in todo.md are marked complete [x], 2) The user's original request has been fully addressed, 3) There are no pending actions or follow-ups required, 4) You've delivered all final outputs and results to the user. IMPORTANT: This is the ONLY way to properly terminate execution. Never use this tool unless ALL tasks are complete and verified. Always ensure you've provided all necessary outputs and references before using this tool. Include relevant attachments when the completion relates to specific files or resources.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Completion message or summary to present to user - should provide clear indication of what was accomplished. Include: 1) Summary of completed tasks, 2) Key deliverables or outputs, 3) Any important notes or next steps, 4) Impact or benefits achieved.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the completion message. Include when: 1) Completion relates to specific files or configurations, 2) User needs to review final outputs, 3) Deliverables are documented in files, 4) Supporting evidence or context is needed. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"web_browser_takeover\",\n      \"description\": \"Request user takeover of browser interaction. Use this tool when: 1) The page requires complex human interaction that automated tools cannot handle, 2) Authentication or verification steps require human input, 3) The page has anti-bot measures that prevent automated access, 4) Complex form filling or navigation is needed, 5) The page requires human verification (CAPTCHA, etc.). IMPORTANT: This tool should be used as a last resort after web-search and crawl-webpage have failed, and when direct browser tools are insufficient. Always provide clear context about why takeover is needed and what actions the user should take.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"text\": {\n            \"type\": \"string\",\n            \"description\": \"Instructions for the user about what actions to take in the browser. Include: 1) Clear explanation of why takeover is needed, 2) Specific steps the user should take, 3) What information to look for or extract, 4) How to indicate when they're done, 5) Any important context about the current page state.\"\n          },\n          \"attachments\": {\n            \"anyOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"type\": \"array\"\n              }\n            ],\n            \"description\": \"(Optional) List of files or URLs to attach to the takeover request. Include when: 1) Screenshots or visual references are needed, 2) Previous search results or crawled content is relevant, 3) Supporting documentation is required. Always use relative paths to /workspace directory.\"\n          }\n        },\n        \"required\": [\n          \"text\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"clear_all\",\n      \"description\": \"Clear all tasks and sections (creates completely empty state).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"confirm\": {\n            \"type\": \"boolean\",\n            \"description\": \"Must be true to confirm clearing everything\"\n          }\n        },\n        \"required\": [\n          \"confirm\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_tasks\",\n      \"description\": \"Create tasks organized by sections. Supports both single section and multi-section batch creation. Creates sections automatically if they don't exist. IMPORTANT: Create tasks in the exact order they will be executed. Each task should represent a single, specific operation that can be completed independently. Break down complex operations into individual, sequential tasks to maintain the one-task-at-a-time execution principle. You MUST specify either 'sections' array OR both 'task_contents' and ('section_title' OR 'section_id').\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"sections\": {\n            \"type\": \"array\",\n            \"description\": \"List of sections with their tasks for batch creation\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"title\": {\n                  \"type\": \"string\",\n                  \"description\": \"Section title (creates if doesn't exist)\"\n                },\n                \"tasks\": {\n                  \"type\": \"array\",\n                  \"description\": \"Task contents for this section\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  },\n                  \"minItems\": 1\n                }\n              },\n              \"required\": [\n                \"title\",\n                \"tasks\"\n              ]\n            }\n          },\n          \"section_title\": {\n            \"type\": \"string\",\n            \"description\": \"Single section title (creates if doesn't exist - use this OR sections array)\"\n          },\n          \"section_id\": {\n            \"type\": \"string\",\n            \"description\": \"Existing section ID (use this OR sections array OR section_title)\"\n          },\n          \"task_contents\": {\n            \"type\": \"array\",\n            \"description\": \"Task contents for single section creation (use with section_title or section_id)\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          }\n        },\n        \"anyOf\": [\n          {\n            \"required\": [\n              \"sections\"\n            ]\n          },\n          {\n            \"required\": [\n              \"task_contents\"\n            ],\n            \"anyOf\": [\n              {\n                \"required\": [\n                  \"section_title\"\n                ]\n              },\n              {\n                \"required\": [\n                  \"section_id\"\n                ]\n              }\n            ]\n          }\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_tasks\",\n      \"description\": \"Delete one or more tasks and/or sections. Can delete tasks by their IDs or sections by their IDs (which will also delete all tasks in those sections).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"task_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Task ID (string) or array of task IDs to delete (optional)\"\n          },\n          \"section_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Section ID (string) or array of section IDs to delete (will also delete all tasks in these sections) (optional)\"\n          },\n          \"confirm\": {\n            \"type\": \"boolean\",\n            \"description\": \"Must be true to confirm deletion of sections (required when deleting sections)\"\n          }\n        },\n        \"anyOf\": [\n          {\n            \"required\": [\n              \"task_ids\"\n            ]\n          },\n          {\n            \"required\": [\n              \"section_ids\",\n              \"confirm\"\n            ]\n          }\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_tasks\",\n      \"description\": \"Update one or more tasks. EFFICIENT BATCHING: Before calling this tool, think about what tasks you have completed and batch them into a single update call. This is more efficient than making multiple consecutive update calls. Always execute tasks in the exact sequence they appear, but batch your updates when possible. Update task status to 'completed' after finishing each task, and consider batching multiple completed tasks into one call rather than updating them individually.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"task_ids\": {\n            \"oneOf\": [\n              {\n                \"type\": \"string\"\n              },\n              {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"string\"\n                },\n                \"minItems\": 1\n              }\n            ],\n            \"description\": \"Task ID (string) or array of task IDs to update. EFFICIENT APPROACH: Batch multiple completed tasks into a single call rather than making multiple consecutive update calls. Always maintain sequential execution order.\"\n          },\n          \"content\": {\n            \"type\": \"string\",\n            \"description\": \"New content for the task(s) (optional)\"\n          },\n          \"status\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"pending\",\n              \"completed\",\n              \"cancelled\"\n            ],\n            \"description\": \"New status for the task(s) (optional). Set to 'completed' for finished tasks. Batch multiple completed tasks when possible.\"\n          },\n          \"section_id\": {\n            \"type\": \"string\",\n            \"description\": \"Section ID to move task(s) to (optional)\"\n          }\n        },\n        \"required\": [\n          \"task_ids\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"view_tasks\",\n      \"description\": \"View all tasks and sections. Use this to see current tasks, check progress, or review completed work. IMPORTANT: This tool helps you identify the next task to execute in the sequential workflow. Always execute tasks in the exact order they appear, completing one task fully before moving to the next. Use this to determine which task is currently pending and should be tackled next.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"check_command_output\",\n      \"description\": \"Check the output of a previously executed command in a tmux session. Use this to monitor the progress or results of non-blocking commands.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the tmux session to check.\"\n          },\n          \"kill_session\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to terminate the tmux session after checking. Set to true when you're done with the command.\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"session_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"execute_command\",\n      \"description\": \"Execute a shell command in the workspace directory. IMPORTANT: Commands are non-blocking by default and run in a tmux session. This is ideal for long-running operations like starting servers or build processes. Uses sessions to maintain state between commands. This tool is essential for running CLI tools, installing packages, and managing system operations.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"command\": {\n            \"type\": \"string\",\n            \"description\": \"The shell command to execute. Use this for running CLI tools, installing packages, or system operations. Commands can be chained using &&, ||, and | operators.\"\n          },\n          \"folder\": {\n            \"type\": \"string\",\n            \"description\": \"Optional relative path to a subdirectory of /workspace where the command should be executed. Example: 'data/pdfs'\"\n          },\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional name of the tmux session to use. Use named sessions for related commands that need to maintain state. Defaults to a random session name.\"\n          },\n          \"blocking\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to wait for the command to complete. Defaults to false for non-blocking execution.\",\n            \"default\": false\n          },\n          \"timeout\": {\n            \"type\": \"integer\",\n            \"description\": \"Optional timeout in seconds for blocking commands. Defaults to 60. Ignored for non-blocking commands.\",\n            \"default\": 60\n          }\n        },\n        \"required\": [\n          \"command\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_commands\",\n      \"description\": \"List all running tmux sessions and their status.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {}\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"terminate_command\",\n      \"description\": \"Terminate a running command by killing its tmux session.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"session_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the tmux session to terminate.\"\n          }\n        },\n        \"required\": [\n          \"session_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_file\",\n      \"description\": \"Create a new file with the provided contents at a given path in the workspace. The path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py)\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be created, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"file_contents\": {\n            \"type\": \"string\",\n            \"description\": \"The content to write to the file\"\n          },\n          \"permissions\": {\n            \"type\": \"string\",\n            \"description\": \"File permissions in octal format (e.g., '644')\",\n            \"default\": \"644\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"file_contents\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_file\",\n      \"description\": \"Delete a file at the given path. The path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py)\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be deleted, relative to /workspace (e.g., 'src/main.py')\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"edit_file\",\n      \"description\": \"Use this tool to make an edit to an existing file.\\n\\nThis will be read by a less intelligent model, which will quickly apply the edit. You should make it clear what the edit is, while also minimizing the unchanged code you write.\\nWhen writing the edit, you should specify each edit in sequence, with the special comment // ... existing code ... to represent unchanged code in between edited lines.\\n\\nFor example:\\n\\n// ... existing code ...\\nFIRST_EDIT\\n// ... existing code ...\\nSECOND_EDIT\\n// ... existing code ...\\nTHIRD_EDIT\\n// ... existing code ...\\n\\nYou should still bias towards repeating as few lines of the original file as possible to convey the change.\\nBut, each edit should contain sufficient context of unchanged lines around the code you're editing to resolve ambiguity.\\nDO NOT omit spans of pre-existing code (or comments) without using the // ... existing code ... comment to indicate its absence. If you omit the existing code comment, the model may inadvertently delete these lines.\\nIf you plan on deleting a section, you must provide context before and after to delete it. If the initial code is ```code \\\\n Block 1 \\\\n Block 2 \\\\n Block 3 \\\\n code```, and you want to remove Block 2, you would output ```// ... existing code ... \\\\n Block 1 \\\\n  Block 3 \\\\n // ... existing code ...```.\\nMake sure it is clear what the edit should be, and where it should be applied.\\nALWAYS make all edits to a file in a single edit_file instead of multiple edit_file calls to the same file. The apply model can handle many distinct edits at once.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"target_file\": {\n            \"type\": \"string\",\n            \"description\": \"The target file to modify\"\n          },\n          \"instructions\": {\n            \"type\": \"string\",\n            \"description\": \"A single sentence written in the first person describing what you're changing. Used to help disambiguate uncertainty in the edit.\"\n          },\n          \"code_edit\": {\n            \"type\": \"string\",\n            \"description\": \"Specify ONLY the precise lines of code that you wish to edit. Use // ... existing code ... for unchanged sections.\"\n          }\n        },\n        \"required\": [\n          \"target_file\",\n          \"instructions\",\n          \"code_edit\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"full_file_rewrite\",\n      \"description\": \"Completely rewrite an existing file with new content. The file path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py). IMPORTANT: Always prefer using edit_file for making changes to code. Only use this tool when edit_file fails or when you need to replace the entire file content.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the file to be rewritten, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"file_contents\": {\n            \"type\": \"string\",\n            \"description\": \"The new content to write to the file, replacing all existing content\"\n          },\n          \"permissions\": {\n            \"type\": \"string\",\n            \"description\": \"File permissions in octal format (e.g., '644')\",\n            \"default\": \"644\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"file_contents\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"str_replace\",\n      \"description\": \"Replace specific text in a file. The file path must be relative to /workspace (e.g., 'src/main.py' for /workspace/src/main.py). IMPORTANT: Prefer using edit_file for faster, shorter edits to avoid repetition. Only use this tool when you need to replace a unique string that appears exactly once in the file and edit_file is not suitable.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the target file, relative to /workspace (e.g., 'src/main.py')\"\n          },\n          \"old_str\": {\n            \"type\": \"string\",\n            \"description\": \"Text to be replaced (must appear exactly once)\"\n          },\n          \"new_str\": {\n            \"type\": \"string\",\n            \"description\": \"Replacement text\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"old_str\",\n          \"new_str\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"deploy\",\n      \"description\": \"Deploy a static website (HTML+CSS+JS) from a directory in the sandbox to Cloudflare Pages. Only use this tool when permanent deployment to a production environment is needed. The directory path must be relative to /workspace. The website will be deployed to {name}.kortix.cloud.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the deployment, will be used in the URL as {name}.kortix.cloud\"\n          },\n          \"directory_path\": {\n            \"type\": \"string\",\n            \"description\": \"Path to the directory containing the static website files to deploy, relative to /workspace (e.g., 'build')\"\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"directory_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"expose_port\",\n      \"description\": \"Expose a port from the agent's sandbox environment to the public internet and get its preview URL. This is essential for making services running in the sandbox accessible to users, such as web applications, APIs, or other network services. The exposed URL can be shared with users to allow them to interact with the sandbox environment.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"port\": {\n            \"type\": \"integer\",\n            \"description\": \"The port number to expose. Must be a valid port number between 1 and 65535.\",\n            \"minimum\": 1,\n            \"maximum\": 65535\n          }\n        },\n        \"required\": [\n          \"port\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"scrape_webpage\",\n      \"description\": \"Extract full text content from multiple webpages in a single operation. IMPORTANT: You should ALWAYS collect multiple relevant URLs from web-search results and scrape them all in a single call for efficiency. This tool saves time by processing multiple pages simultaneously rather than one at a time. The extracted text includes the main content of each page without HTML markup.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"urls\": {\n            \"type\": \"string\",\n            \"description\": \"Multiple URLs to scrape, separated by commas. You should ALWAYS include several URLs when possible for efficiency. Example: 'https://example.com/page1,https://example.com/page2,https://example.com/page3'\"\n          }\n        },\n        \"required\": [\n          \"urls\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"web_search\",\n      \"description\": \"Search the web for up-to-date information on a specific topic using the Tavily API. This tool allows you to gather real-time information from the internet to answer user queries, research topics, validate facts, and find recent developments. Results include titles, URLs, and publication dates. Use this tool for discovering relevant web pages before potentially crawling them for complete content.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"query\": {\n            \"type\": \"string\",\n            \"description\": \"The search query to find relevant web pages. Be specific and include key terms to improve search accuracy. For best results, use natural language questions or keyword combinations that precisely describe what you're looking for.\"\n          },\n          \"num_results\": {\n            \"type\": \"integer\",\n            \"description\": \"The number of search results to return. Increase for more comprehensive research or decrease for focused, high-relevance results.\",\n            \"default\": 20\n          }\n        },\n        \"required\": [\n          \"query\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"see_image\",\n      \"description\": \"Allows the agent to 'see' an image file located in the /workspace directory or from a URL. Provide either a relative path to a local image or the URL to an image. The image will be compressed before sending to reduce token usage. The image content will be made available in the next turn's context.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\",\n            \"description\": \"Either a relative path to the image file within the /workspace directory (e.g., 'screenshots/image.png') or a URL to an image (e.g., 'https://example.com/image.jpg'). Supported formats: JPG, PNG, GIF, WEBP. Max size: 10MB.\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"image_edit_or_generate\",\n      \"description\": \"Generate a new image from a prompt, or edit an existing image (no mask support) using OpenAI GPT Image 1 via OpenAI SDK. Stores the result in the thread context.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"mode\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"generate\",\n              \"edit\"\n            ],\n            \"description\": \"'generate' to create a new image from a prompt, 'edit' to edit an existing image.\"\n          },\n          \"prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Text prompt describing the desired image or edit.\"\n          },\n          \"image_path\": {\n            \"type\": \"string\",\n            \"description\": \"(edit mode only) Path to the image file to edit. Can be: 1) Relative path to /workspace (e.g., 'generated_image_abc123.png'), or 2) Full URL (e.g., 'https://example.com/image.png'). Required when mode='edit'.\"\n          }\n        },\n        \"required\": [\n          \"mode\",\n          \"prompt\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_presentation_outline\",\n      \"description\": \"Create a structured outline for a presentation with slide titles and descriptions. This tool helps plan the overall structure and flow of a presentation before creating the actual slides. The final presentation will use FLAT DESIGN principles with clean typography, solid colors, and no gradients/shadows/animations. Standard PowerPoint dimensions (1920x1080, 16:9 aspect ratio) will be maintained.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"title\": {\n            \"type\": \"string\",\n            \"description\": \"The main title of the presentation\"\n          },\n          \"subtitle\": {\n            \"type\": \"string\",\n            \"description\": \"Optional subtitle or tagline for the presentation\"\n          },\n          \"slides\": {\n            \"type\": \"array\",\n            \"description\": \"Array of slide outlines with title and description\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"title\": {\n                  \"type\": \"string\",\n                  \"description\": \"The title of the slide\"\n                },\n                \"description\": {\n                  \"type\": \"string\",\n                  \"description\": \"Brief description of the slide content and purpose\"\n                },\n                \"notes\": {\n                  \"type\": \"string\",\n                  \"description\": \"Optional speaker notes or additional context\"\n                }\n              },\n              \"required\": [\n                \"title\",\n                \"description\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"title\",\n          \"slides\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_presentation\",\n      \"description\": \"\\n            Create a professional presentation using structured JSON format.\\n            The presentation will be saved as JSON and can be previewed in the browser or exported to PPTX.\\n            \\n            IMPORTANT: Generate a structured JSON with specific layout types. Each slide must have:\\n            - layout: One of 'title', 'title-bullets', 'title-content', 'two-column', 'image-text', \\n              'quote', 'section', 'blank', 'hero-image', 'image-grid', 'comparison', 'timeline', 'stats'\\n            - content: Object with fields specific to that layout type\\n            \\n            Images can be specified as:\\n            - URLs (will be downloaded and embedded)\\n            - Unsplash search terms using \\\"unsplash:keyword\\\" format\\n            - Local file paths in the workspace\\n            \\n            The tool will handle all formatting and ensure consistency across preview and export.\\n            \\n            THEME SELECTION: Instead of a fixed theme, choose a theme that best fits the content:\\n            - 'corporate-blue': Professional blue theme for business presentations\\n            - 'modern-purple': Modern purple/pink gradient for tech/startup\\n            - 'minimal-mono': Black and white minimalist design\\n            - 'ocean-teal': Calming teal and aqua colors\\n            - 'sunset-warm': Warm orange and red tones\\n            - 'forest-green': Natural green palette\\n            - 'midnight-dark': Dark mode with bright accents\\n            - 'pastel-soft': Soft pastel colors\\n            - 'bold-contrast': High contrast bold colors\\n            - 'elegant-gold': Sophisticated gold and navy\\n            \\n            Select the theme that best matches the presentation's content and purpose.\\n            \",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"presentation_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the presentation (used for file naming)\"\n          },\n          \"title\": {\n            \"type\": \"string\",\n            \"description\": \"Main title of the presentation\"\n          },\n          \"subtitle\": {\n            \"type\": \"string\",\n            \"description\": \"Optional subtitle or tagline\"\n          },\n          \"theme\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"corporate-blue\",\n              \"modern-purple\",\n              \"minimal-mono\",\n              \"ocean-teal\",\n              \"sunset-warm\",\n              \"forest-green\",\n              \"midnight-dark\",\n              \"pastel-soft\",\n              \"bold-contrast\",\n              \"elegant-gold\"\n            ],\n            \"description\": \"Visual theme - choose based on content and purpose\"\n          },\n          \"slides\": {\n            \"type\": \"array\",\n            \"description\": \"Array of slide objects with layout and content\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"layout\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"title\",\n                    \"title-bullets\",\n                    \"title-content\",\n                    \"two-column\",\n                    \"image-text\",\n                    \"quote\",\n                    \"section\",\n                    \"blank\",\n                    \"hero-image\",\n                    \"image-grid\",\n                    \"comparison\",\n                    \"timeline\",\n                    \"stats\"\n                  ],\n                  \"description\": \"Layout type for the slide\"\n                },\n                \"content\": {\n                  \"type\": \"object\",\n                  \"description\": \"Content object specific to the layout type\",\n                  \"properties\": {\n                    \"title\": {\n                      \"type\": \"string\"\n                    },\n                    \"subtitle\": {\n                      \"type\": \"string\"\n                    },\n                    \"bullets\": {\n                      \"type\": \"array\",\n                      \"items\": {\n                        \"type\": \"string\"\n                      }\n                    },\n                    \"text\": {\n                      \"type\": \"string\"\n                    },\n                    \"left_content\": {\n                      \"type\": \"object\"\n                    },\n                    \"right_content\": {\n                      \"type\": \"object\"\n                    },\n                    \"image\": {\n                      \"type\": \"object\",\n                      \"properties\": {\n                        \"url\": {\n                          \"type\": \"string\"\n                        },\n                        \"alt\": {\n                          \"type\": \"string\"\n                        },\n                        \"position\": {\n                          \"type\": \"string\",\n                          \"enum\": [\n                            \"left\",\n                            \"right\",\n                            \"center\",\n                            \"background\"\n                          ]\n                        }\n                      }\n                    },\n                    \"quote\": {\n                      \"type\": \"string\"\n                    },\n                    \"author\": {\n                      \"type\": \"string\"\n                    },\n                    \"notes\": {\n                      \"type\": \"string\"\n                    }\n                  }\n                }\n              },\n              \"required\": [\n                \"layout\",\n                \"content\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"presentation_name\",\n          \"title\",\n          \"slides\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"export_presentation\",\n      \"description\": \"Export a JSON presentation to PPTX format with professional formatting.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"presentation_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the presentation to export\"\n          },\n          \"format\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"pptx\"\n            ],\n            \"description\": \"Export format (currently only PPTX supported)\"\n          }\n        },\n        \"required\": [\n          \"presentation_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"analyze_sheet\",\n      \"description\": \"Simple statistics and optional group_by; can export CSV.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"target_columns\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"group_by\": {\n            \"type\": \"string\"\n          },\n          \"aggregations\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\",\n              \"enum\": [\n                \"count\",\n                \"sum\",\n                \"avg\",\n                \"min\",\n                \"max\"\n              ]\n            }\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_sheet\",\n      \"description\": \"Create a new CSV/XLSX with optional headers/rows.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"headers\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"rows\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"array\",\n              \"items\": {\n                \"type\": \"string\"\n              }\n            }\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"overwrite\": {\n            \"type\": \"boolean\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"format_sheet\",\n      \"description\": \"Style and format sheet cells (bold headers, auto widths, optional conditional formatting). XLSX only.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"bold_headers\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"auto_width\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"apply_banding\": {\n            \"type\": \"boolean\",\n            \"default\": true\n          },\n          \"conditional_format\": {\n            \"type\": \"object\",\n            \"properties\": {\n              \"column\": {\n                \"type\": \"string\"\n              },\n              \"min_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFEFEB\"\n              },\n              \"mid_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFD7D2\"\n              },\n              \"max_color\": {\n                \"type\": \"string\",\n                \"default\": \"FFA39E\"\n              }\n            }\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_sheet\",\n      \"description\": \"Modify existing cells, rows, or columns (insert/delete/update).\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"operations\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"type\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"update_cell\",\n                    \"update_row\",\n                    \"insert_row\",\n                    \"delete_row\",\n                    \"insert_column\",\n                    \"delete_column\"\n                  ]\n                },\n                \"row_index\": {\n                  \"type\": \"integer\"\n                },\n                \"column\": {\n                  \"type\": \"string\"\n                },\n                \"column_index\": {\n                  \"type\": \"integer\"\n                },\n                \"values\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  }\n                },\n                \"value\": {\n                  \"type\": \"string\"\n                }\n              },\n              \"required\": [\n                \"type\"\n              ]\n            }\n          },\n          \"save_as\": {\n            \"type\": \"string\",\n            \"description\": \"Optional path to save result (.csv or .xlsx)\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"operations\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"view_sheet\",\n      \"description\": \"Read headers, types, and sample rows; optional CSV export.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"max_rows\": {\n            \"type\": \"integer\",\n            \"default\": 100\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\"\n          }\n        },\n        \"required\": [\n          \"file_path\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"visualize_sheet\",\n      \"description\": \"Generate charts (bar, line, pie, scatter) and save to XLSX. Also optionally export chart data to CSV for Google Sheets.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"file_path\": {\n            \"type\": \"string\"\n          },\n          \"sheet_name\": {\n            \"type\": \"string\",\n            \"nullable\": true\n          },\n          \"chart_type\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"bar\",\n              \"line\",\n              \"pie\",\n              \"scatter\"\n            ],\n            \"default\": \"bar\"\n          },\n          \"x_column\": {\n            \"type\": \"string\"\n          },\n          \"y_columns\": {\n            \"type\": \"array\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"save_as\": {\n            \"type\": \"string\"\n          },\n          \"export_csv_path\": {\n            \"type\": \"string\",\n            \"description\": \"Optional CSV path for the chart dataset (x + y columns)\"\n          }\n        },\n        \"required\": [\n          \"file_path\",\n          \"x_column\",\n          \"y_columns\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"build_project\",\n      \"description\": \"Run production build for the project.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"project_name\": {\n            \"type\": \"string\"\n          },\n          \"package_manager\": {\n            \"type\": \"string\",\n            \"default\": \"pnpm\"\n          }\n        },\n        \"required\": [\n          \"project_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_project_structure\",\n      \"description\": \"Get the file structure of a web project, showing important files and directories.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"project_name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the project directory to examine\"\n          },\n          \"max_depth\": {\n            \"type\": \"integer\",\n            \"description\": \"Maximum depth to traverse (default: 3)\",\n            \"default\": 3\n          }\n        },\n        \"required\": [\n          \"project_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"execute_data_provider_call\",\n      \"description\": \"Execute a call to a specific data provider endpoint\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"service_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the API service (e.g., 'linkedin')\"\n          },\n          \"route\": {\n            \"type\": \"string\",\n            \"description\": \"The key of the endpoint to call\"\n          },\n          \"payload\": {\n            \"type\": \"object\",\n            \"description\": \"The payload to send with the API call\"\n          }\n        },\n        \"required\": [\n          \"service_name\",\n          \"route\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_data_provider_endpoints\",\n      \"description\": \"Get available endpoints for a specific data provider\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"service_name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the data provider (e.g., 'linkedin', 'twitter', 'zillow', 'amazon', 'yahoo_finance')\"\n          }\n        },\n        \"required\": [\n          \"service_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_current_agent_config\",\n      \"description\": \"Get the current configuration of the agent being edited. Use this to check what's already configured before making updates.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_agent\",\n      \"description\": \"Update the agent's configuration including name, description, system prompt, tools, and MCP servers. Call this whenever the user wants to modify any aspect of the agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"The name of the agent. Should be descriptive and indicate the agent's purpose.\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"A brief description of what the agent does and its capabilities.\"\n          },\n          \"system_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"The system instructions that define the agent's behavior, expertise, and approach. This should be comprehensive and well-structured.\"\n          },\n          \"agentpress_tools\": {\n            \"type\": \"object\",\n            \"description\": \"Configuration for AgentPress tools. Each key is a tool name, and the value is a boolean indicating if the tool is enabled.\",\n            \"additionalProperties\": {\n              \"type\": \"boolean\"\n            }\n          },\n          \"configured_mcps\": {\n            \"type\": \"array\",\n            \"description\": \"List of configured MCP servers for external integrations.\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"name\": {\n                  \"type\": \"string\"\n                },\n                \"qualifiedName\": {\n                  \"type\": \"string\"\n                },\n                \"config\": {\n                  \"type\": \"object\"\n                },\n                \"enabledTools\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"type\": \"string\"\n                  }\n                }\n              }\n            }\n          },\n          \"avatar\": {\n            \"type\": \"string\",\n            \"description\": \"Emoji to use as the agent's avatar.\"\n          },\n          \"avatar_color\": {\n            \"type\": \"string\",\n            \"description\": \"Hex color code for the agent's avatar background.\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"discover_user_mcp_servers\",\n      \"description\": \"Discover available MCP tools for a specific Composio profile. Use this to see what MCP tools are available for a connected profile.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The profile ID from the Composio credential profile\"\n          }\n        },\n        \"required\": [\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_app_details\",\n      \"description\": \"Get detailed information about a specific Composio toolkit, including available tools and authentication requirements.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"The toolkit slug to get details for (e.g., 'github', 'linear', 'slack')\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"search_mcp_servers\",\n      \"description\": \"Search for Composio toolkits based on user requirements. Use this when the user wants to add MCP tools to their agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"query\": {\n            \"type\": \"string\",\n            \"description\": \"Search query for finding relevant Composio toolkits (e.g., 'linear', 'github', 'database', 'search')\"\n          },\n          \"limit\": {\n            \"type\": \"integer\",\n            \"description\": \"Maximum number of toolkits to return (default: 10)\",\n            \"default\": 10\n          }\n        },\n        \"required\": [\n          \"query\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"configure_profile_for_agent\",\n      \"description\": \"Configure a connected credential profile to be used by the agent with selected tools. Use this after the profile is connected and you want to add it to the agent.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the connected credential profile\"\n          },\n          \"enabled_tools\": {\n            \"type\": \"array\",\n            \"description\": \"List of tool names to enable for this profile\",\n            \"items\": {\n              \"type\": \"string\"\n            }\n          },\n          \"display_name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional custom display name for this configuration in the agent\"\n          }\n        },\n        \"required\": [\n          \"profile_id\",\n          \"enabled_tools\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_credential_profile\",\n      \"description\": \"Create a new Composio credential profile for a specific toolkit. This will create the integration and return an authentication link that the user needs to visit to connect their account.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"The toolkit slug to create the profile for (e.g., 'github', 'linear', 'slack')\"\n          },\n          \"profile_name\": {\n            \"type\": \"string\",\n            \"description\": \"A name for this credential profile (e.g., 'Personal GitHub', 'Work Slack')\"\n          },\n          \"display_name\": {\n            \"type\": \"string\",\n            \"description\": \"Display name for the profile (defaults to profile_name if not provided)\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\",\n          \"profile_name\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_credential_profile\",\n      \"description\": \"Delete a credential profile that is no longer needed. This will also remove it from any agent configurations.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"The ID of the credential profile to delete\"\n          }\n        },\n        \"required\": [\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_credential_profiles\",\n      \"description\": \"Get all existing Composio credential profiles for the current user. Use this to show the user their available profiles.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Optional filter to show only profiles for a specific toolkit\"\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"activate_workflow\",\n      \"description\": \"Activate or deactivate a workflow/playbook. Only active workflows can be executed.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to activate/deactivate\"\n          },\n          \"active\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to activate (true) or deactivate (false) the workflow\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_workflow\",\n      \"description\": \"Create a new workflow/playbook. This stores a Start node with a single child that contains config.playbook { template, variables }.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the workflow\"\n          },\n          \"template\": {\n            \"type\": \"string\",\n            \"description\": \"Workflow/playbook instructions text. Use {{variable}} tokens.\"\n          },\n          \"variables\": {\n            \"type\": \"array\",\n            \"description\": \"Optional variable specs; all treated as strings currently\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"key\": {\n                  \"type\": \"string\"\n                },\n                \"label\": {\n                  \"type\": \"string\"\n                },\n                \"required\": {\n                  \"type\": \"boolean\",\n                  \"default\": true\n                }\n              },\n              \"required\": [\n                \"key\"\n              ]\n            },\n            \"default\": []\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"Optional short description (auto-summarized from template if omitted)\"\n          },\n          \"is_default\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether this should be the default workflow\",\n            \"default\": false\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"template\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_workflow\",\n      \"description\": \"Delete a workflow/playbook from the agent. This action cannot be undone.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to delete\"\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_workflows\",\n      \"description\": \"Get all workflows/playbooks for the current agent. Use this to see what workflows are already configured.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"include_steps\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include detailed step information for each workflow\",\n            \"default\": true\n          }\n        },\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"update_workflow\",\n      \"description\": \"Update an existing workflow or playbook. For playbooks, the first child under Start should contain config.playbook { template, variables }.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to update\"\n          },\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"New name for the workflow\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"New description for the workflow\"\n          },\n          \"trigger_phrase\": {\n            \"type\": \"string\",\n            \"description\": \"New trigger phrase for the workflow\"\n          },\n          \"is_default\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether this workflow should be the default workflow\"\n          },\n          \"status\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"draft\",\n              \"active\",\n              \"inactive\"\n            ],\n            \"description\": \"Status of the workflow\"\n          },\n          \"validate_tools\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to validate tool names against available tools when updating steps\",\n            \"default\": true\n          },\n          \"steps\": {\n            \"type\": \"array\",\n            \"description\": \"New steps. For playbooks, follow the Start + Execute Playbook structure with config.playbook.\",\n            \"items\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"name\": {\n                  \"type\": \"string\"\n                },\n                \"description\": {\n                  \"type\": \"string\"\n                },\n                \"type\": {\n                  \"type\": \"string\",\n                  \"enum\": [\n                    \"instruction\",\n                    \"tool\",\n                    \"condition\"\n                  ],\n                  \"default\": \"instruction\"\n                },\n                \"config\": {\n                  \"type\": \"object\",\n                  \"additionalProperties\": true\n                },\n                \"conditions\": {\n                  \"type\": \"object\",\n                  \"additionalProperties\": true\n                },\n                \"order\": {\n                  \"type\": \"integer\"\n                },\n                \"children\": {\n                  \"type\": \"array\",\n                  \"items\": {\n                    \"$ref\": \"#\"\n                  }\n                }\n              },\n              \"required\": [\n                \"name\",\n                \"order\"\n              ]\n            }\n          }\n        },\n        \"required\": [\n          \"workflow_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_event_trigger\",\n      \"description\": \"Create a Composio event-based trigger for this agent. First list apps and triggers, then pass the chosen trigger slug, profile_id, and trigger_config. Optionally route to a workflow.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"slug\": {\n            \"type\": \"string\",\n            \"description\": \"Trigger type slug, e.g. 'GMAIL_NEW_GMAIL_MESSAGE'\"\n          },\n          \"profile_id\": {\n            \"type\": \"string\",\n            \"description\": \"Composio profile_id to use (must be connected)\"\n          },\n          \"trigger_config\": {\n            \"type\": \"object\",\n            \"description\": \"Trigger configuration object per trigger schema\",\n            \"additionalProperties\": true\n          },\n          \"route\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"agent\",\n              \"workflow\"\n            ],\n            \"default\": \"agent\",\n            \"description\": \"Execute agent directly or run a workflow\"\n          },\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Optional friendly name for the trigger\"\n          },\n          \"agent_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Prompt to pass to the agent when route is 'agent'\"\n          },\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"Workflow ID when route is 'workflow'\"\n          },\n          \"workflow_input\": {\n            \"type\": \"object\",\n            \"description\": \"Workflow input variables when route is 'workflow'\",\n            \"additionalProperties\": true\n          },\n          \"connected_account_id\": {\n            \"type\": \"string\",\n            \"description\": \"Connected account id; if omitted we try to derive from profile\"\n          }\n        },\n        \"required\": [\n          \"slug\",\n          \"profile_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"create_scheduled_trigger\",\n      \"description\": \"Create a scheduled trigger for the agent to execute workflows or direct agent runs using cron expressions. This allows the agent to run automatically at specified times.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name of the scheduled trigger. Should be descriptive of when/why it runs.\"\n          },\n          \"description\": {\n            \"type\": \"string\",\n            \"description\": \"Description of what this trigger does and when it runs.\"\n          },\n          \"cron_expression\": {\n            \"type\": \"string\",\n            \"description\": \"Cron expression defining when to run (e.g., '0 9 * * *' for daily at 9am, '*/30 * * * *' for every 30 minutes)\"\n          },\n          \"execution_type\": {\n            \"type\": \"string\",\n            \"enum\": [\n              \"workflow\",\n              \"agent\"\n            ],\n            \"description\": \"Whether to execute a workflow or run the agent directly\",\n            \"default\": \"agent\"\n          },\n          \"workflow_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the workflow to execute (required if execution_type is 'workflow')\"\n          },\n          \"workflow_input\": {\n            \"type\": \"object\",\n            \"description\": \"Input data to pass to the workflow (optional, only for workflow execution)\",\n            \"additionalProperties\": true\n          },\n          \"agent_prompt\": {\n            \"type\": \"string\",\n            \"description\": \"Prompt to send to the agent when triggered (required if execution_type is 'agent')\"\n          }\n        },\n        \"required\": [\n          \"name\",\n          \"cron_expression\",\n          \"execution_type\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"delete_scheduled_trigger\",\n      \"description\": \"Delete a scheduled trigger. The agent will no longer run automatically at the scheduled time.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"trigger_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the trigger to delete\"\n          }\n        },\n        \"required\": [\n          \"trigger_id\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"get_scheduled_triggers\",\n      \"description\": \"Get all scheduled triggers for the current agent. Shows when the agent will run automatically.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_app_event_triggers\",\n      \"description\": \"List available triggers for a given app/toolkit slug. Includes slug, name, description, type, instructions, config, and payload schema.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Toolkit slug, e.g. 'gmail'\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_event_profiles\",\n      \"description\": \"List connected Composio profiles for a toolkit. Use this to get profile_id and connected_account_id before creating a trigger.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"toolkit_slug\": {\n            \"type\": \"string\",\n            \"description\": \"Toolkit slug, e.g. 'gmail'\"\n          }\n        },\n        \"required\": [\n          \"toolkit_slug\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"list_event_trigger_apps\",\n      \"description\": \"List apps (toolkits) that have available event-based triggers via Composio. Returns slug, name, and logo.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {},\n        \"required\": []\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"toggle_scheduled_trigger\",\n      \"description\": \"Enable or disable a scheduled trigger. Disabled triggers won't run until re-enabled.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"trigger_id\": {\n            \"type\": \"string\",\n            \"description\": \"ID of the trigger to toggle\"\n          },\n          \"is_active\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to enable (true) or disable (false) the trigger\"\n          }\n        },\n        \"required\": [\n          \"trigger_id\",\n          \"is_active\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_act\",\n      \"description\": \"Perform any browser action using natural language description. CRITICAL: This tool automatically provides a screenshot with every action. For data entry actions (filling forms, entering text, selecting options), you MUST review the provided screenshot to verify that displayed values exactly match what was intended. Report mismatches immediately.\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"action\": {\n            \"type\": \"string\",\n            \"description\": \"The action to perform. Examples: 'click the login button', 'fill in the email field with test@example.com', 'scroll down to see more content', 'select option 2 from the dropdown', 'press Enter', 'go back', 'wait 5 seconds', 'click at coordinates 100,200', 'drag the file icon to the drop zone', 'select United States from the country dropdown'\"\n          },\n          \"variables\": {\n            \"type\": \"object\",\n            \"description\": \"Variables to use in the action. Variables in the action string are referenced using %variable_name%. These variables are NOT shared with LLM providers for security.\",\n            \"additionalProperties\": {\n              \"type\": \"string\"\n            },\n            \"default\": {}\n          },\n          \"iframes\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include iframe content in the action. Set to true if the target element is inside an iframe.\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"action\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_extract_content\",\n      \"description\": \"Extract structured content from the current page using Stagehand\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"instruction\": {\n            \"type\": \"string\",\n            \"description\": \"What content to extract (e.g., 'extract all product prices', 'get the main heading', 'extract apartment listings with address and price')\"\n          },\n          \"selector\": {\n            \"type\": \"string\",\n            \"description\": \"Optional XPath selector to reduce extraction scope to a specific element. Useful for reducing input tokens and increasing accuracy.\",\n            \"default\": null\n          },\n          \"iframes\": {\n            \"type\": \"boolean\",\n            \"description\": \"Whether to include iframe content in the extraction. Set to true if the target content is inside an iframe.\",\n            \"default\": true\n          }\n        },\n        \"required\": [\n          \"instruction\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_navigate_to\",\n      \"description\": \"Navigate to a specific url\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"url\": {\n            \"type\": \"string\",\n            \"description\": \"The url to navigate to\"\n          }\n        },\n        \"required\": [\n          \"url\"\n        ]\n      }\n    }\n  },\n  {\n    \"type\": \"function\",\n    \"function\": {\n      \"name\": \"browser_screenshot\",\n      \"description\": \"Take a screenshot of the current page\",\n      \"parameters\": {\n        \"type\": \"object\",\n        \"properties\": {\n          \"name\": {\n            \"type\": \"string\",\n            \"description\": \"Name for the screenshot\",\n            \"default\": \"screenshot\"\n          }\n        }\n      }\n    }\n  }\n]\n```\n\nWhen using the tools:\n- Use the exact function names from the JSON schema above\n- Include all required parameters as specified in the schema\n- Format complex data (objects, arrays) as JSON strings within the parameter tags\n- Boolean values should be \"true\" or \"false\" (lowercase)\n\n\nUsage Examples:\n\nexpand_message:\n\n        <!-- Example 1: Expand a message that was truncated in the previous conversation -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">ecde3a4c-c7dc-4776-ae5c-8209517c5576</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Expand a message to create reports or analyze truncated data -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">f47ac10b-58cc-4372-a567-0e02b2c3d479</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Expand a message when you need the full content for analysis -->\n        <function_calls>\n        <invoke name=\"expand_message\">\n        <parameter name=\"message_id\">550e8400-e29b-41d4-a716-446655440000</parameter>\n        </invoke>\n        </function_calls>\n        \n\nask:\n\n        <function_calls>\n        <invoke name=\"ask\">\n        <parameter name=\"text\">I'm planning to bake the chocolate cake for your birthday party. The recipe mentions \"rich frosting\" but doesn't specify what type. Could you clarify your preferences? For example:\n1. Would you prefer buttercream or cream cheese frosting?\n2. Do you want any specific flavor added to the frosting (vanilla, coffee, etc.)?\n3. Should I add any decorative toppings like sprinkles or fruit?\n4. Do you have any dietary restrictions I should be aware of?\n\nThis information will help me make sure the cake meets your expectations for the celebration.</parameter>\n        <parameter name=\"attachments\">recipes/chocolate_cake.txt,photos/cake_examples.jpg</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncomplete:\n\n        <function_calls>\n        <invoke name=\"complete\">\n        <parameter name=\"text\">I have successfully completed all tasks for your project. Here's what was accomplished:\n1. Created the web application with modern UI components\n2. Implemented user authentication and database integration\n3. Deployed the application to production\n4. Created comprehensive documentation\n\nAll deliverables are attached for your review.</parameter>\n        <parameter name=\"attachments\">app/src/main.js,docs/README.md,deployment-config.yaml</parameter>\n        </invoke>\n        </function_calls>\n        \n\nweb_browser_takeover:\n\n        <function_calls>\n        <invoke name=\"web_browser_takeover\">\n        <parameter name=\"text\">I've encountered a CAPTCHA verification on the page that I can't solve automatically. Could you help me out?\n\nHere's what I need you to do:\n1. Solve the CAPTCHA puzzle that's currently displayed\n2. Let me know once you've completed it\n3. I'll then continue with the automated process\n\nIf you encounter any issues or need to take additional steps, please let me know. Thanks for your help!</parameter>\n        </invoke>\n        </function_calls>\n        \n\nclear_all:\n\n        <function_calls>\n        <invoke name=\"clear_all\">\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_tasks:\n\n        # Batch creation across multiple sections:\n        <function_calls>\n        <invoke name=\"create_tasks\">\n        <parameter name=\"sections\">[\n            {\n                \"title\": \"Setup & Planning\", \n                \"tasks\": [\"Research requirements\", \"Create project plan\"]\n            },\n            {\n                \"title\": \"Development\", \n                \"tasks\": [\"Setup environment\", \"Write code\", \"Add tests\"]\n            },\n            {\n                \"title\": \"Deployment\", \n                \"tasks\": [\"Deploy to staging\", \"Run tests\", \"Deploy to production\"]\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Simple single section creation:\n        <function_calls>\n        <invoke name=\"create_tasks\">\n        <parameter name=\"section_title\">Bug Fixes</parameter>\n        <parameter name=\"task_contents\">[\"Fix login issue\", \"Update error handling\"]</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_tasks:\n\n        # Delete single task:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">task-uuid-here</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete multiple tasks:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\"]</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete single section (and all its tasks):\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"section_ids\">section-uuid-here</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete multiple sections (and all their tasks):\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"section_ids\">[\"section-id-1\", \"section-id-2\"]</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Delete both tasks and sections:\n        <function_calls>\n        <invoke name=\"delete_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\"]</parameter>\n        <parameter name=\"section_ids\">[\"section-id-1\"]</parameter>\n        <parameter name=\"confirm\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nupdate_tasks:\n\n        # Update single task (when only one task is completed):\n        <function_calls>\n        <invoke name=\"update_tasks\">\n        <parameter name=\"task_ids\">task-uuid-here</parameter>\n        <parameter name=\"status\">completed</parameter>\n        </invoke>\n        </function_calls>\n        \n        # Update multiple tasks (EFFICIENT: batch multiple completed tasks):\n        <function_calls>\n        <invoke name=\"update_tasks\">\n        <parameter name=\"task_ids\">[\"task-id-1\", \"task-id-2\", \"task-id-3\"]</parameter>\n        <parameter name=\"status\">completed</parameter>\n        </invoke>\n        </function_calls>\n        \n\nview_tasks:\n\n        <function_calls>\n        <invoke name=\"view_tasks\">\n        </invoke>\n        </function_calls>\n        \n\ncheck_command_output:\n\n        <function_calls>\n        <invoke name=\"check_command_output\">\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n        \n        <!-- Example 2: Check final output and kill session -->\n        <function_calls>\n        <invoke name=\"check_command_output\">\n        <parameter name=\"session_name\">build_process</parameter>\n        <parameter name=\"kill_session\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexecute_command:\n\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm run dev</parameter>\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Running in Specific Directory -->\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm run build</parameter>\n        <parameter name=\"folder\">frontend</parameter>\n        <parameter name=\"session_name\">build_process</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Blocking command (wait for completion) -->\n        <function_calls>\n        <invoke name=\"execute_command\">\n        <parameter name=\"command\">npm install</parameter>\n        <parameter name=\"blocking\">true</parameter>\n        <parameter name=\"timeout\">300</parameter>\n        </invoke>\n        </function_calls>\n        \n\nlist_commands:\n\n        <function_calls>\n        <invoke name=\"list_commands\">\n        </invoke>\n        </function_calls>\n        \n\nterminate_command:\n\n        <function_calls>\n        <invoke name=\"terminate_command\">\n        <parameter name=\"session_name\">dev_server</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_file:\n\n        <function_calls>\n        <invoke name=\"create_file\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"file_contents\">\n        # This is the file content\n        def main():\n            print(\"Hello, World!\")\n        \n        if __name__ == \"__main__\":\n            main()\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_file:\n\n        <function_calls>\n        <invoke name=\"delete_file\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        </invoke>\n        </function_calls>\n        \n\nedit_file:\n\n        <!-- Example: Mark multiple scattered tasks as complete in a todo list -->\n        <function_calls>\n        <invoke name=\"edit_file\">\n        <parameter name=\"target_file\">todo.md</parameter>\n        <parameter name=\"instructions\">I am marking the research and setup tasks as complete in my todo list.</parameter>\n        <parameter name=\"code_edit\">\n// ... existing code ...\n- [x] Research topic A\n- [ ] Research topic B\n- [x] Research topic C\n// ... existing code ...\n- [x] Setup database\n- [x] Configure server\n// ... existing code ...\n        </parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example: Add error handling and logging to a function -->\n        <function_calls>\n        <invoke name=\"edit_file\">\n        <parameter name=\"target_file\">src/main.py</parameter>\n        <parameter name=\"instructions\">I am adding error handling and logging to the user authentication function</parameter>\n        <parameter name=\"code_edit\">\n// ... existing imports ...\nfrom my_app.logging import logger\nfrom my_app.exceptions import DatabaseError\n// ... existing code ...\ndef authenticate_user(username, password):\n    try:\n        user = get_user(username)\n        if user and verify_password(password, user.password_hash):\n            return user\n        return None\n    except DatabaseError as e:\n        logger.error(f\"Database error during authentication: {e}\")\n        return None\n    except Exception as e:\n        logger.error(f\"Unexpected error during authentication: {e}\")\n        return None\n// ... existing code ...\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\nfull_file_rewrite:\n\n        <function_calls>\n        <invoke name=\"full_file_rewrite\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"file_contents\">\n        This completely replaces the entire file content.\n        Use when making major changes to a file or when the changes\n        are too extensive for str-replace.\n        All previous content will be lost and replaced with this text.\n        </parameter>\n        </invoke>\n        </function_calls>\n        \n\nstr_replace:\n\n        <function_calls>\n        <invoke name=\"str_replace\">\n        <parameter name=\"file_path\">src/main.py</parameter>\n        <parameter name=\"old_str\">text to replace (must appear exactly once in the file)</parameter>\n        <parameter name=\"new_str\">replacement text that will be inserted instead</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndeploy:\n\n        <!-- \n        IMPORTANT: Only use this tool when:\n        1. The user explicitly requests permanent deployment to production\n        2. You have a complete, ready-to-deploy directory \n        \n        NOTE: If the same name is used, it will redeploy to the same project as before\n        -->\n\n        <function_calls>\n        <invoke name=\"deploy\">\n        <parameter name=\"name\">my-site</parameter>\n        <parameter name=\"directory_path\">website</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexpose_port:\n\n        <!-- Example 1: Expose a web server running on port 8000 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">8000</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 2: Expose an API service running on port 3000 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">3000</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example 3: Expose a development server running on port 5173 -->\n        <function_calls>\n        <invoke name=\"expose_port\">\n        <parameter name=\"port\">5173</parameter>\n        </invoke>\n        </function_calls>\n        \n\nscrape_webpage:\n\n        <function_calls>\n        <invoke name=\"scrape_webpage\">\n        <parameter name=\"urls\">https://www.kortix.ai/,https://github.com/kortix-ai/suna</parameter>\n        </invoke>\n        </function_calls>\n        \n\nweb_search:\n\n        <function_calls>\n        <invoke name=\"web_search\">\n        <parameter name=\"query\">what is Kortix AI and what are they building?</parameter>\n        <parameter name=\"num_results\">20</parameter>\n        </invoke>\n        </function_calls>\n        \n        <!-- Another search example -->\n        <function_calls>\n        <invoke name=\"web_search\">\n        <parameter name=\"query\">latest AI research on transformer models</parameter>\n        <parameter name=\"num_results\">20</parameter>\n        </invoke>\n        </function_calls>\n        \n\nsee_image:\n\n        <!-- Example: Request to see a local image named 'diagram.png' inside the 'docs' folder -->\n        <function_calls>\n        <invoke name=\"see_image\">\n        <parameter name=\"file_path\">docs/diagram.png</parameter>\n        </invoke>\n        </function_calls>\n\n        <!-- Example: Request to see an image from a URL -->\n        <function_calls>\n        <invoke name=\"see_image\">\n        <parameter name=\"file_path\">https://example.com/image.jpg</parameter>\n        </invoke>\n        </function_calls>\n        \n\nimage_edit_or_generate:\n\n        Generate mode example (new image):\n        <function_calls>\n        <invoke name=\"image_edit_or_generate\">\n        <parameter name=\"mode\">generate</parameter>\n        <parameter name=\"prompt\">A futuristic cityscape at sunset</parameter>\n        </invoke>\n        </function_calls>\n        \n        Edit mode example (modifying existing):\n        <function_calls>\n        <invoke name=\"image_edit_or_generate\">\n        <parameter name=\"mode\">edit</parameter>\n        <parameter name=\"prompt\">Add a red hat to the person in the image</parameter>\n        <parameter name=\"image_path\">generated_image_abc123.png</parameter>\n        </invoke>\n        </function_calls>\n        \n        Multi-turn workflow (follow-up edits):\n        1. User: \"Create a logo\" \u2192 generate mode\n        2. User: \"Make it more colorful\" \u2192 edit mode (automatic)\n        3. User: \"Add text to it\" \u2192 edit mode (automatic)\n        \n\ncreate_presentation_outline:\n\n        <function_calls>\n        <invoke name=\"create_presentation_outline\">\n        <parameter name=\"title\">The Future of AI</parameter>\n        <parameter name=\"subtitle\">Transforming how we work, create, and connect</parameter>\n        <parameter name=\"slides\">[\n            {\n                \"title\": \"The Future of AI\",\n                \"description\": \"Title slide with bold typography and clean flat design, setting a professional tone for the presentation.\",\n                \"notes\": \"Open with confidence. Use a solid color background with high contrast text. Keep the design minimal and impactful.\"\n            },\n            {\n                \"title\": \"Revolutionary Technology\",\n                \"description\": \"Content slide showcasing key AI capabilities with clean typography and simple bullet points.\",\n                \"notes\": \"Highlight the four key areas: neural networks, automation, creativity, and connectivity. Use flat icons or simple graphics.\"\n            },\n            {\n                \"title\": \"The Question\",\n                \"description\": \"Minimal slide with large, bold text asking a thought-provoking question.\",\n                \"notes\": \"Pause for emphasis. Use a contrasting flat color background. Let the question sink in before continuing.\"\n            },\n            {\n                \"title\": \"Real-World Impact\",\n                \"description\": \"Showcase concrete AI applications across industries with clean layout and organized content sections.\",\n                \"notes\": \"Present examples from healthcare, transportation, climate, and education using flat design principles.\"\n            },\n            {\n                \"title\": \"The Road Ahead\",\n                \"description\": \"Inspirational slide with a powerful quote and forward-looking message about human potential.\",\n                \"notes\": \"Quote: 'The best way to predict the future is to invent it.' Use clean typography and solid colors.\"\n            },\n            {\n                \"title\": \"Thank You\",\n                \"description\": \"Clean, minimal closing slide inviting questions, using flat design with good typography.\",\n                \"notes\": \"Keep it simple. Use a solid background color with contrasting text. Create space for audience engagement.\"\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_presentation:\n\n        <function_calls>\n        <invoke name=\"create_presentation\">\n        <parameter name=\"presentation_name\">company_overview</parameter>\n        <parameter name=\"title\">Company Overview 2024</parameter>\n        <parameter name=\"subtitle\">Innovation Through Technology</parameter>\n        <parameter name=\"theme\">modern-purple</parameter>\n        <parameter name=\"slides\">[\n            {\n                \"layout\": \"hero-image\",\n                \"content\": {\n                    \"title\": \"Welcome to the Future\",\n                    \"subtitle\": \"Where Innovation Meets Excellence\",\n                    \"image\": {\n                        \"url\": \"unsplash:technology office\",\n                        \"position\": \"background\"\n                    }\n                }\n            },\n            {\n                \"layout\": \"stats\",\n                \"content\": {\n                    \"title\": \"Our Impact in Numbers\",\n                    \"stats\": [\n                        {\"value\": \"50K+\", \"label\": \"Active Users\"},\n                        {\"value\": \"$2.5M\", \"label\": \"Revenue\"},\n                        {\"value\": \"98%\", \"label\": \"Satisfaction\"},\n                        {\"value\": \"15\", \"label\": \"Countries\"}\n                    ]\n                }\n            },\n            {\n                \"layout\": \"image-text\",\n                \"content\": {\n                    \"title\": \"Our Mission\",\n                    \"text\": \"We empower businesses with cutting-edge AI technology to transform their operations, enhance productivity, and unlock new possibilities. Our platform combines powerful automation with human creativity.\",\n                    \"image\": {\n                        \"url\": \"unsplash:artificial intelligence\",\n                        \"position\": \"right\"\n                    }\n                }\n            },\n            {\n                \"layout\": \"image-grid\",\n                \"content\": {\n                    \"title\": \"Our Products in Action\",\n                    \"images\": [\n                        {\"url\": \"unsplash:dashboard analytics\", \"caption\": \"Real-time Analytics\"},\n                        {\"url\": \"unsplash:team collaboration\", \"caption\": \"Team Collaboration\"},\n                        {\"url\": \"unsplash:mobile app\", \"caption\": \"Mobile Experience\"},\n                        {\"url\": \"unsplash:data visualization\", \"caption\": \"Data Insights\"}\n                    ]\n                }\n            },\n            {\n                \"layout\": \"title-bullets\",\n                \"content\": {\n                    \"title\": \"Key Features\",\n                    \"bullets\": [\n                        \"AI-powered automation for repetitive tasks\",\n                        \"Real-time collaboration and communication\",\n                        \"Advanced analytics and reporting\",\n                        \"Enterprise-grade security and compliance\",\n                        \"24/7 customer support and training\"\n                    ]\n                }\n            },\n            {\n                \"layout\": \"quote\",\n                \"content\": {\n                    \"quote\": \"This platform has transformed how we work. We've saved 40% of our time and increased productivity by 60%.\",\n                    \"author\": \"Sarah Johnson, CTO at TechCorp\"\n                }\n            },\n            {\n                \"layout\": \"section\",\n                \"content\": {\n                    \"title\": \"Let's Build Together\",\n                    \"subtitle\": \"Start Your Journey Today\"\n                }\n            },\n            {\n                \"layout\": \"title\",\n                \"content\": {\n                    \"title\": \"Thank You\",\n                    \"subtitle\": \"Questions? Let's Connect!\",\n                    \"notes\": \"Contact us at hello@company.com\"\n                }\n            }\n        ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\nanalyze_sheet:\n\n        <function_calls>\n        <invoke name=\"analyze_sheet\">\n          <parameter name=\"file_path\">data/sales.csv</parameter>\n          <parameter name=\"target_columns\">[\"revenue\"]</parameter>\n          <parameter name=\"group_by\">region</parameter>\n          <parameter name=\"export_csv_path\">data/sales_summary.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_sheet:\n\n        <function_calls>\n        <invoke name=\"create_sheet\">\n          <parameter name=\"file_path\">data/sample.csv</parameter>\n          <parameter name=\"headers\">[\"region\",\"revenue\"]</parameter>\n          <parameter name=\"rows\">[[\"NA\",100],[\"EU\",120]]</parameter>\n          <parameter name=\"overwrite\">true</parameter>\n        </invoke>\n        </function_calls>\n    \n\nformat_sheet:\n\n        <function_calls>\n        <invoke name=\"format_sheet\">\n          <parameter name=\"file_path\">reports/sales_chart.xlsx</parameter>\n          <parameter name=\"sheet_name\">Data</parameter>\n          <parameter name=\"bold_headers\">true</parameter>\n          <parameter name=\"conditional_format\">{\"column\":\"revenue\"}</parameter>\n        </invoke>\n        </function_calls>\n    \n\nupdate_sheet:\n\n        <function_calls>\n        <invoke name=\"update_sheet\">\n          <parameter name=\"file_path\">employee_details.xlsx</parameter>\n          <parameter name=\"sheet_name\">Employee Database</parameter>\n          <parameter name=\"operations\">[\n            {\"type\":\"insert_row\",\"row_index\":6,\"values\":[\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]},\n            {\"type\":\"insert_row\",\"row_index\":7,\"values\":[\"SUMMARY CALCULATIONS\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]},\n            {\"type\":\"insert_row\",\"row_index\":8,\"values\":[\"Total Employees:\",\"=COUNTA(A2:A100)\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\"]}\n          ]</parameter>\n        </invoke>\n        </function_calls>\n    \n\nview_sheet:\n\n        <function_calls>\n        <invoke name=\"view_sheet\">\n          <parameter name=\"file_path\">reports/sales.xlsx</parameter>\n          <parameter name=\"sheet_name\">Sheet1</parameter>\n          <parameter name=\"max_rows\">50</parameter>\n          <parameter name=\"export_csv_path\">reports/sales_preview.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\nvisualize_sheet:\n\n        <function_calls>\n        <invoke name=\"visualize_sheet\">\n          <parameter name=\"file_path\">reports/sales.csv</parameter>\n          <parameter name=\"x_column\">region</parameter>\n          <parameter name=\"y_columns\">[\"revenue\"]</parameter>\n          <parameter name=\"chart_type\">bar</parameter>\n          <parameter name=\"save_as\">reports/sales_chart.xlsx</parameter>\n          <parameter name=\"export_csv_path\">reports/sales_chart_data.csv</parameter>\n        </invoke>\n        </function_calls>\n    \n\nbuild_project:\n\n        <!-- Build a Next.js project for production -->\n        <function_calls>\n        <invoke name=\"build_project\">\n        <parameter name=\"project_name\">my-app</parameter>\n        <parameter name=\"package_manager\">pnpm</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_project_structure:\n\n        <!-- Get structure of a project -->\n        <function_calls>\n        <invoke name=\"get_project_structure\">\n        <parameter name=\"project_name\">my-app</parameter>\n        </invoke>\n        </function_calls>\n        \n\nexecute_data_provider_call:\n\n        <!-- \n        The execute-data-provider-call tool makes a request to a specific data provider endpoint.\n        Use this tool when you need to call an data provider endpoint with specific parameters.\n        The route must be a valid endpoint key obtained from get-data-provider-endpoints tool!!\n        -->\n        \n        <!-- Example to call linkedIn service with the specific route person -->\n        <function_calls>\n        <invoke name=\"execute_data_provider_call\">\n        <parameter name=\"service_name\">linkedin</parameter>\n        <parameter name=\"route\">person</parameter>\n        <parameter name=\"payload\">{\"link\": \"https://www.linkedin.com/in/johndoe/\"}</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_data_provider_endpoints:\n\n<!-- \nThe get-data-provider-endpoints tool returns available endpoints for a specific data provider.\nUse this tool when you need to discover what endpoints are available.\n-->\n\n<!-- Example to get LinkedIn API endpoints -->\n<function_calls>\n<invoke name=\"get_data_provider_endpoints\">\n<parameter name=\"service_name\">linkedin</parameter>\n</invoke>\n</function_calls>\n        \n\nget_current_agent_config:\n\n        <function_calls>\n        <invoke name=\"get_current_agent_config\">\n        </invoke>\n        </function_calls>\n        \n\nupdate_agent:\n\n        <function_calls>\n        <invoke name=\"update_agent\">\n        <parameter name=\"name\">Research Assistant</parameter>\n        <parameter name=\"description\">An AI assistant specialized in conducting research and providing comprehensive analysis</parameter>\n        <parameter name=\"system_prompt\">You are a research assistant with expertise in gathering, analyzing, and synthesizing information. Your approach is thorough and methodical...</parameter>\n                        <parameter name=\"agentpress_tools\">{\"web_search_tool\": true, \"sb_files_tool\": true, \"sb_shell_tool\": false}</parameter>\n        <parameter name=\"avatar\">\ud83d\udd2c</parameter>\n        <parameter name=\"avatar_color\">#4F46E5</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndiscover_user_mcp_servers:\n\n        <function_calls>\n        <invoke name=\"discover_user_mcp_servers\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_app_details:\n\n        <function_calls>\n        <invoke name=\"get_app_details\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        </invoke>\n        </function_calls>\n        \n\nsearch_mcp_servers:\n\n        <function_calls>\n        <invoke name=\"search_mcp_servers\">\n        <parameter name=\"query\">linear</parameter>\n        <parameter name=\"limit\">5</parameter>\n        </invoke>\n        </function_calls>\n        \n\nconfigure_profile_for_agent:\n\n        <function_calls>\n        <invoke name=\"configure_profile_for_agent\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        <parameter name=\"enabled_tools\">[\"create_issue\", \"list_repositories\", \"get_user\"]</parameter>\n        <parameter name=\"display_name\">Personal GitHub Integration</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_credential_profile:\n\n        <function_calls>\n        <invoke name=\"create_credential_profile\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        <parameter name=\"profile_name\">Personal GitHub</parameter>\n        <parameter name=\"display_name\">My Personal GitHub Account</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_credential_profile:\n\n        <function_calls>\n        <invoke name=\"delete_credential_profile\">\n        <parameter name=\"profile_id\">profile-uuid-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_credential_profiles:\n\n        <function_calls>\n        <invoke name=\"get_credential_profiles\">\n        <parameter name=\"toolkit_slug\">github</parameter>\n        </invoke>\n        </function_calls>\n        \n\nactivate_workflow:\n\n        <function_calls>\n        <invoke name=\"activate_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"active\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_workflow:\n\n        <function_calls>\n        <invoke name=\"create_workflow\">\n        <parameter name=\"name\">Company Research Playbook</parameter>\n        <parameter name=\"template\">Research companies and update sheets using {{sheet_id}} and {{limit}}</parameter>\n        <parameter name=\"variables\">[{\"key\":\"sheet_id\",\"label\":\"Sheet ID\",\"required\":true},{\"key\":\"limit\",\"label\":\"Row Limit\",\"required\":false}]</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_workflow:\n\n        <function_calls>\n        <invoke name=\"delete_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_workflows:\n\n        <function_calls>\n        <invoke name=\"get_workflows\">\n        <parameter name=\"include_steps\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nupdate_workflow:\n\n        <function_calls>\n        <invoke name=\"update_workflow\">\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"name\">Updated Research Workflow</parameter>\n        <parameter name=\"status\">active</parameter>\n        </invoke>\n        </function_calls>\n        \n\ncreate_event_trigger:\n\n        <function_calls>\n        <invoke name=\"list_event_trigger_apps\"></invoke>\n        <invoke name=\"list_app_event_triggers\"><parameter name=\"toolkit_slug\">gmail</parameter></invoke>\n        <invoke name=\"list_event_profiles\"><parameter name=\"toolkit_slug\">gmail</parameter></invoke>\n        <invoke name=\"create_event_trigger\">\n          <parameter name=\"slug\">GMAIL_NEW_GMAIL_MESSAGE</parameter>\n          <parameter name=\"profile_id\">profile_123</parameter>\n          <parameter name=\"trigger_config\">{\"interval\": 1, \"userId\": \"me\", \"labelIds\": \"INBOX\"}</parameter>\n          <parameter name=\"route\">agent</parameter>\n          <parameter name=\"agent_prompt\">Read this</parameter>\n        </invoke>\n        </function_calls>\n    \n\ncreate_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"create_scheduled_trigger\">\n        <parameter name=\"name\">Daily Report Generation</parameter>\n        <parameter name=\"description\">Generates daily reports every morning at 9 AM</parameter>\n        <parameter name=\"cron_expression\">0 9 * * *</parameter>\n        <parameter name=\"execution_type\">workflow</parameter>\n        <parameter name=\"workflow_id\">workflow-123</parameter>\n        <parameter name=\"workflow_input\">{\"report_type\": \"daily\", \"include_charts\": true}</parameter>\n        </invoke>\n        </function_calls>\n        \n\ndelete_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"delete_scheduled_trigger\">\n        <parameter name=\"trigger_id\">trigger-123</parameter>\n        </invoke>\n        </function_calls>\n        \n\nget_scheduled_triggers:\n\n        <function_calls>\n        <invoke name=\"get_scheduled_triggers\">\n        </invoke>\n        </function_calls>\n        \n\nlist_app_event_triggers:\n\n        <function_calls>\n        <invoke name=\"list_app_event_triggers\">\n        <parameter name=\"toolkit_slug\">gmail</parameter>\n        </invoke>\n        </function_calls>\n    \n\nlist_event_profiles:\n\n        <function_calls>\n        <invoke name=\"list_event_profiles\">\n        <parameter name=\"toolkit_slug\">gmail</parameter>\n        </invoke>\n        </function_calls>\n    \n\nlist_event_trigger_apps:\n\n        <function_calls>\n        <invoke name=\"list_event_trigger_apps\"></invoke>\n        </function_calls>\n    \n\ntoggle_scheduled_trigger:\n\n        <function_calls>\n        <invoke name=\"toggle_scheduled_trigger\">\n        <parameter name=\"trigger_id\">trigger-123</parameter>\n        <parameter name=\"is_active\">false</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_act:\n\n        <function_calls>\n        <invoke name=\"browser_act\">\n        <parameter name=\"action\">fill in the login form with %username% and %password%</parameter>\n        <parameter name=\"variables\">{\"username\": \"john.doe\", \"password\": \"secret123\"}</parameter>\n        <parameter name=\"iframes\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_extract_content:\n\n        <function_calls>\n        <invoke name=\"browser_extract_content\">\n        <parameter name=\"instruction\">extract all product names and prices from the main product list</parameter>\n        <parameter name=\"selector\">//div[@class='product-list']</parameter>\n        <parameter name=\"iframes\">true</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_navigate_to:\n\n        <function_calls>\n        <invoke name=\"browser_navigate_to\">\n        <parameter name=\"url\">https://example.com</parameter>\n        </invoke>\n        </function_calls>\n        \n\nbrowser_screenshot:\n\n        <function_calls>\n        <invoke name=\"browser_screenshot\">\n        <parameter name=\"name\">page_screenshot</parameter>\n        </invoke>\n        </function_calls>\n        \n"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"activate_workflow\", \"xml_tag_name\": \"activate-workflow\", \"tool_call_id\": null, \"result\": {\"success\": true, \"output\": {\"message\": \"Workflow 'Comprehensive Company Research Playbook' activated successfully\", \"workflow_id\": \"d3fd2202-6256-42e3-98d3-ff4a4fc562cb\", \"status\": \"active\"}, \"error\": null}}}",
      "message_id": "876b93b3-25ed-4cad-a16a-446af90e313d"
    },
    {
      "role": "user",
      "content": "continue",
      "message_id": "79a118c1-b22c-41d8-8a05-7a64be7c7063"
    },
    {
      "role": "assistant",
      "content": "Great! The workflow is now active. Let me start by updating the first task as complete and begin the actual research process. I'll start with a comprehensive company research example to demonstrate the full workflow.\n\n<function_calls>\n<invoke name=\"update_tasks\">\n<parameter name=\"task_ids\">[\"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\"]</parameter>\n<parameter name=\"status\">completed</para... (truncated)\n\nmessage_id \"bf621916-1013-4aec-bf9a-5b93e25e0a5a\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "bf621916-1013-4aec-bf9a-5b93e25e0a5a"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"update_tasks\", \"xml_tag_name\": \"update-tasks\", \"tool_call_id\": null, \"result\": {\"success\": true, \"output\": {\"sections\": [{\"id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\", \"title\": \"Research Setup & Planning\", \"tasks\": [{\"id\": \"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\", \"content\": \"Set up company research workflow with variables\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"content\": \"Configure data sources and search parameters\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"20b5484e-fe54-4236-8f5c-e71a738a26da\", \"content\": \"Establish research methodology and sources\", \"status\": \"pending\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}]}, {\"id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\", \"title\": \"Company Fundamentals Research\", \"tasks\": [{\"id\": \"b08bc6af-f98c-46a4-b965-8c96c963ea33\", \"content\": \"Research basic company information and overview\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"d97ef9f3-3d4a-47b0-8ae6-b8137c11ea32\", \"content\": \"Gather recent news and announcements\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"a93178e5-a6d7-4c2c-8a30-50bb78987d36\", \"content\": \"Research funding history and investment rounds\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"7ac8c5a2-8cc0-4cfa-9202-44812e7a6a33\", \"content\": \"Analyze financial performance and metrics\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}]}, {\"id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\", \"title\": \"Leadership & Personnel Analysis\", \"tasks\": [{\"id\": \"a1ae7fc4-5abd-4197-b783-319c874de228\", \"content\": \"Research CEO and executive team backgrounds\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"238940d8-d6cc-45e7-83b3-a6377137201c\", \"content\": \"Analyze key personnel LinkedIn profiles\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"ddbe449e-33a5-4675-862e-43775a0b2730\", \"content\": \"Research board members and advisors\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"643e917a-088c-4140-8426-cafb7e600faa\", \"content\": \"Identify leadership changes and transitions\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}]}, {\"id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\", \"title\": \"Market Position Analysis\", \"tasks\": [{\"id\": \"f4e90132-5b18-4b3c-9018-39c911414e9b\", \"content\": \"Research competitive landscape\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"454b1d2b-4da1-4c14-8f6e-31073a789702\", \"content\": \"Analyze market share and positioning\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"92282e56-bee1-44c2-a4ba-7679472ea5d1\", \"content\": \"Identify key competitors and differentiators\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"aa1fe7aa-c4ec-44c5-85d3-2cac70fe88c7\", \"content\": \"Research industry trends and market dynamics\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}]}, {\"id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\", \"title\": \"Insights & Profile Creation\", \"tasks\": [{\"id\": \"f78993ad-b1b9-4fc8-8896-5616c3da35c1\", \"content\": \"Synthesize all research findings\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"5e1421e0-55b6-4bdc-9d20-11c67ffbd6da\", \"content\": \"Create comprehensive company profile\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"1060c31b-66f7-4f69-b5aa-b7f0b8623731\", \"content\": \"Generate actionable insights and recommendations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"32f09439-effe-407d-86b0-084f86c0c990\", \"content\": \"Format final deliverable with citations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}]}], \"total_tasks\": 19, \"total_sections\": 5}, \"error\": null}}}",
      "message_id": "1d3c8600-fc02-4cff-900d-401d35778ffc"
    },
    {
      "role": "assistant",
      "content": "Now let me demonstrate the complete research workflow by researching **Stripe** as an example company. I'll execute each task systematically to show you how this comprehensive research process works.\n\n<function_calls>\n<invoke name=\"update_tasks\">\n<parameter name=\"task_ids\">[\"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"20b5484e-fe54-4236-8f5c-e71a738a26da\"]</parameter>\n<parameter name=\"... (truncated)\n\nmessage_id \"7c9dc8a4-641f-4db1-92d4-8fd39752031e\"\nUse expand-message tool to see contents",
      "tool_calls": null,
      "message_id": "7c9dc8a4-641f-4db1-92d4-8fd39752031e"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"update_tasks\", \"xml_tag_name\": \"update-tasks\", \"tool_call_id\": null, \"result\": {\"success\": true, \"output\": {\"sections\": [{\"id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\", \"title\": \"Research Setup & Planning\", \"tasks\": [{\"id\": \"f2aa48a5-a2f1-46f9-9b47-d88ff214babd\", \"content\": \"Set up company research workflow with variables\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"16a5f529-7daf-461f-8898-f4f7abaf7e4e\", \"content\": \"Configure data sources and search parameters\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}, {\"id\": \"20b5484e-fe54-4236-8f5c-e71a738a26da\", \"content\": \"Establish research methodology and sources\", \"status\": \"completed\", \"section_id\": \"07ab39ff-77f7-4543-a877-2c718b6c16dd\"}]}, {\"id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\", \"title\": \"Company Fundamentals Research\", \"tasks\": [{\"id\": \"b08bc6af-f98c-46a4-b965-8c96c963ea33\", \"content\": \"Research basic company information and overview\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"d97ef9f3-3d4a-47b0-8ae6-b8137c11ea32\", \"content\": \"Gather recent news and announcements\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"a93178e5-a6d7-4c2c-8a30-50bb78987d36\", \"content\": \"Research funding history and investment rounds\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}, {\"id\": \"7ac8c5a2-8cc0-4cfa-9202-44812e7a6a33\", \"content\": \"Analyze financial performance and metrics\", \"status\": \"pending\", \"section_id\": \"2ac36597-48f6-4c3c-80ee-717c229558df\"}]}, {\"id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\", \"title\": \"Leadership & Personnel Analysis\", \"tasks\": [{\"id\": \"a1ae7fc4-5abd-4197-b783-319c874de228\", \"content\": \"Research CEO and executive team backgrounds\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"238940d8-d6cc-45e7-83b3-a6377137201c\", \"content\": \"Analyze key personnel LinkedIn profiles\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"ddbe449e-33a5-4675-862e-43775a0b2730\", \"content\": \"Research board members and advisors\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}, {\"id\": \"643e917a-088c-4140-8426-cafb7e600faa\", \"content\": \"Identify leadership changes and transitions\", \"status\": \"pending\", \"section_id\": \"2eb7c73c-ed49-4fcf-bbb7-aa735ad25986\"}]}, {\"id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\", \"title\": \"Market Position Analysis\", \"tasks\": [{\"id\": \"f4e90132-5b18-4b3c-9018-39c911414e9b\", \"content\": \"Research competitive landscape\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"454b1d2b-4da1-4c14-8f6e-31073a789702\", \"content\": \"Analyze market share and positioning\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"92282e56-bee1-44c2-a4ba-7679472ea5d1\", \"content\": \"Identify key competitors and differentiators\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}, {\"id\": \"aa1fe7aa-c4ec-44c5-85d3-2cac70fe88c7\", \"content\": \"Research industry trends and market dynamics\", \"status\": \"pending\", \"section_id\": \"99ae871e-e573-4a91-9ab9-bbc5035ae6f0\"}]}, {\"id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\", \"title\": \"Insights & Profile Creation\", \"tasks\": [{\"id\": \"f78993ad-b1b9-4fc8-8896-5616c3da35c1\", \"content\": \"Synthesize all research findings\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"5e1421e0-55b6-4bdc-9d20-11c67ffbd6da\", \"content\": \"Create comprehensive company profile\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"1060c31b-66f7-4f69-b5aa-b7f0b8623731\", \"content\": \"Generate actionable insights and recommendations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}, {\"id\": \"32f09439-effe-407d-86b0-084f86c0c990\", \"content\": \"Format final deliverable with citations\", \"status\": \"pending\", \"section_id\": \"f2c20a61-0cb6-449a-8b68-c6204c26105d\"}]}], \"total_tasks\": 19, \"total_sections\": 5}, \"error\": null}}}",
      "message_id": "27f25d61-6912-4755-a09a-e615738aa325"
    },
    {
      "role": "assistant",
      "content": "Now let me begin the actual research on Stripe. I'll start with basic company information and work through each section systematically.\n\n<function_calls>\n<invoke name=\"web_search\">\n<parameter name=\"query\">Stripe company overview business model revenue 2024 2025</parameter>\n<parameter name=\"num_results\">15</parameter>\n</invoke>\n</function_calls>",
      "tool_calls": null,
      "message_id": "4ab32962-1f27-4c6c-b680-bbd0c6d4ebb9"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"web_search\", \"xml_tag_name\": \"web-search\", \"tool_call_id\": null, \"result\": {\"success\": true, \"output\": {\"query\": \"Stripe company overview business model revenue 2024 2025\", \"follow_up_questions\": null, \"answer\": \"Stripe is a leading financial infrastructure company that processes payments for millions of businesses worldwide, handling over $1.4 trillion in total payment volume in 2024, representing approximately 1.3% of global GDP. The company operates on a commission-based business model, typically charging 2.9% plus 30 cents per transaction for domestic cards, with custom pricing available for enterprise clients. Stripe achieved profitability in 2024 and generated an estimated $18-19 billion in revenue, with projections suggesting continued growth into 2025. The platform serves over 1.3 million active websites globally across 46+ countries, supporting 135+ currencies and payment methods, with more than 100 enterprise customers each processing over $1 billion annually through Stripe. Key revenue drivers include their core payment processing services, subscription billing tools (which reached a $500 million revenue run rate and serve over 300,000 companies), and emerging products like Stripe Connect, Stripe Issuing, and AI-powered automation tools. The company is currently valued at approximately $70-91.5 billion and has recently made strategic acquisitions including Bridge, a stablecoin platform purchased for $1.1 billion in 2025, signaling expansion into cryptocurrency infrastructure and positioning for future growth in digital financial services.\", \"images\": [\"https://helplama.com/wp-content/uploads/2023/06/stripe-growth.jpg\", \"https://businessmodelanalyst.com/wp-content/uploads/2021/07/Stripe-Business-Model-Canvas.jpeg\", \"https://lumosbusiness.com/wp-content/uploads/2022/10/Stripe_Business_Model_Canvas.jpg\", \"https://images.ctfassets.net/kepwgvmwgwnt/sWy772Vs4SZRWREoDAnhF/8b5c6b555540daa1db3367aba2013700/stripe-mrr-dashboard.png\", \"https://kinsta.com/wp-content/uploads/2022/09/Stripe-Revenue-per-Year.png\"], \"results\": [{\"url\": \"https://coinlaw.io/stripe-statistics/\", \"title\": \"Stripe Statistics 2025: Usage, Revenue, and Market Share\", \"content\": \"Stripe generated $19.4 billion in revenue in 2025, reflecting a 17% year-over-year growth from 2024.\\n   Subscription tools contributed $3.4 billion in revenue, solidifying their role as a core business driver.\\n   The company posted a 10.6% net profit margin in 2025, maintaining profitability amid fintech competition.\\n   Annual recurring revenue (ARR) exceeded $6.1 billion in 2025, fueled by enterprise-scale adoption. [...] Stripe\\u2019s AI automation tools reduced manual workflows by 26% in 2025, improving operational efficiency for businesses.\\n   The platform\\u2019s custom checkout APIs saw 41% growth in enterprise adoption, enabling highly personalized user experiences.\\n   Stripe Issuing grew by 58%, processing over $13.4 billion in virtual and physical card transactions in 2025.\\n   Stripe Connect generated $340 million in new revenue streams in 2025, driven by platform economy expansion. [...] In 2025, Stripe processed $1.05 trillion in total payment volume, marking a 16% year-over-year increase.\\n   The platform handled $58 billion in cross-border payments, supported by smarter FX and localized payout tools.\\n   Apple Pay and Google Pay transactions grew 31%, now making up 38% of Stripe\\u2019s mobile payment volume.\\n   Stripe\\u2019s fraud system blocked $2.3 billion in fraudulent activity in 2025, showcasing AI-powered risk prevention.\", \"score\": 0.94761914, \"raw_content\": null}, {\"url\": \"https://assets.stripeassets.com/fzn2n1nzq965/2pt3yIHthraqR1KwXgr98U/df10795aac0205789956c89e0dfc4f1a/Stripe-annual-letter-2024.pdf\", \"title\": \"[PDF] Stripe-annual-letter-2024.pdf\", \"content\": \"February 27, 2025 Dear Stripe community, Businesses on Stripe generated $1.4 trillion in total payment volume in 2024, up 38% from the prior year, and reaching a scale equivalent to around 1.3% of global GDP. We attribute this year\\u2019s rapid growth in part to our long-standing investments in building machine learning and artificial intelligence into our products. These bets continue to pay off, increasing revenue for existing customers, encouraging more businesses to switch to Stripe, and helping [...] new companies reach significant scale unprecedentedly quickly\\u2014more on all of those points later. Future growth rates will fluctuate (2024 was an unusually good year), but we\\u2019re as enthusiastic as ever about the long-term trends in the internet economy. Stripe was profitable in 2024, and we expect to be so in 2025 and beyond. Durable profitability allows us to plow back much of our operating earnings into research and development. In each of the last six years, Stripe has reinvested a much [...] right by both and , recognizing its power and maturity. As we forecast in our , Stripe\\u2019s Revenue and Finance Automation Suite, with Billing at its core, has now passed a $500 million revenue run rate. Billing is being used by more than 300,000 companies, manages nearly 200 million active subscriptions, and is emerging as the revenue engine of the AI era. The businesses on Stripe span every chromosome of the economic genome, from top corporate leaders (half of the Fortune 100 uses Stripe) to\", \"score\": 0.9377661, \"raw_content\": null}, {\"url\": \"https://capitaloneshopping.com/research/stripe-statistics/\", \"title\": \"Stripe Statistics (2025): Revenue, Market Share & Growth ...\", \"content\": \"All Reports\\n\\nHighlights. 1.31 million active websites globally use Stripe, with an average of 1,000 companies joining Stripe each day.\\n\\n In 2024, Stripe processed a global payment volume of over $1.4 trillion.\\n As of May 2025, Stripe is valued at $91.5 billion.\\n In 2024, Stripe had an estimated net revenue of $5.1 billion.\\n Stripe holds a 20.8% market share in payment processing software.\\n Stripe has raised a lifetime total of $9.4 billion in investments.\\n\\n## Stripe User Statistics [...] PayPal services are used by 11.9 million websites worldwide; PayPal services are used by 805% more websites than Stripe.\\n In 2024, Stripe competitors Square and Adyen were valued at $40.7 billion and $61.9 billion respectively, compared to Stripe\\u2019s $91.5 billion valuation.\\n In 2024, Stripe recorded a net revenue of $5.1 billion 136% more than Adyen\\u2019s net revenue of $2.16 billion.\\n Square has raised $9.4 billion in investments and Adyen only $266 million.\\n\\n## Stripe Black Friday Statistics [...] | Ranking | Company | 2024 Revenue  (in billions) |\\n| 1 | Walmart | $681.0 |\\n| 2 | Amazon | $650.3 |\\n| 3 | Apple | $400.4 |\\n| 4 | Samsung | $219.6 |\\n| 5 | Walgreens | $151.9 |\\n| 6 | AliExpress | $130.4 |\\n| 7 | Target | $106.6 |\\n| 8 | Chewy | $11.86 |\\n| 9 | Wayfair | $11.85 |\\n| 10 | ASOS | $3.81 |\\n\\n## Stripe User Growth Statistics\\n\\nStripe is one of the fastest-growing online payment processors.\", \"score\": 0.903098, \"raw_content\": null}, {\"url\": \"https://coinlaw.io/paypal-vs-stripe-statistics/\", \"title\": \"PayPal vs. Stripe Statistics 2025: User Base, Fees & Global ...\", \"content\": \"PayPal\\u2019s total revenue reached $32.1 billion in 2025, growing 8.6% YoY.\\n Stripe\\u2019s projected revenue for 2025 stands at $18.9 billion, up 11.2% from $17 billion in 2024.\\n Stripe posted its first-ever quarterly profit in Q1 2025, driven by enterprise growth and infrastructure cost optimizations.\\n PayPal\\u2019s operating margin held steady at 19.4%, while Stripe grew to 12.2% in 2025.\\n Stripe Treasury became profitable for the first time, handling $3.1 billion in business balances and payouts. [...] PayPal\\u2019s revenue from Braintree grew 13.5%, contributing $7.4 billion to its topline.\\n Stripe\\u2019s Revenue Intelligence tools brought in $1.2 billion in new upsell-driven income in 2025.\\n PayPal saw net income of $5.3 billion in 2025, while Stripe posted an estimated net income of $2.6 billion.\\n PayPal\\u2018s recurring revenue streams made up 42% of total revenue, indicating successful diversification efforts. [...] PayPal\\u2018s peer-to-peer (P2P) volume dropped slightly to $315 billion in 2025, as Venmo\\u2019s growth plateaued.\\n Stripe\\u2019s recurring billing volume reached $190 billion in 2025, fueled by its Billing and Invoicing APIs.\\n The average transaction value on PayPal was $84.60 per payment in 2025.\\n The average Stripe transaction came in higher at $98.10, due to its B2B-heavy customer base.\\n PayPal processed over 5.4 billion transactions in Q1\\u2013Q2 of 2025 alone.\", \"score\": 0.8967681, \"raw_content\": null}, {\"url\": \"https://www.chargeflow.io/blog/stripe-statistics\", \"title\": \"Stripe Stats for 2025: The Most Accurate Stripe Statistics ...\", \"content\": \"1.  Stripe generated $1.4 trillion in total payment volume in 2024, up 38% from the prior year.\\n2.  Merchants in 50 countries use Stripe to accept and process payments.\\n3.  More than 100 Stripe customers process $1 billion or more yearly (40 in 2020).\\n4.  4,776,230 million websites have used Stripe to accept payment.\\n5.  Stripe is currently valued at $70 billion, making it one of the most valuable privately-held companies on the planet. [...] ### Stripe Black Friday Cyber Monday Statistics\\n\\nStripe\\u2019s BFCM stats are always a thing for the books and 2024 was no different. Businesses processed more than $31 billion on Stripe from Black Friday through Cyber Monday, the company proclaimed. That\\u2019s a notable uptick from $18.6 billion in 2023. [...] Image 31\\n\\nStripe Revenue Statistics\\n-------------------------\\n\\nStripe's business model is charging a commission for every transaction processed, typically at a standard rate of 2.9% + 30 cents (for domestic cards) of the transaction. However, they offer custom pricing options for enterprise clients. Stripe Radar, a fraud protection tool, charges 5 to 7 cents for each screened transaction.\", \"score\": 0.8452414, \"raw_content\": null}, {\"url\": \"https://sacra.com/c/stripe/valuation/\", \"title\": \"Stripe: $91.50B valuation [2025 - Sacra\", \"content\": \"| Bull \\ud83d\\ude80 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $23B | $30B | $40B | $53B | $71B |\\n| Growth | 25.00% | 27.69% | 30.00% | 32.31% | 33.81% | 35% |\\n\\nOur bear, base, and bull cases for Stripe reflect varying scenarios of competitive pressure, market penetration, and ecosystem expansion success [...] | Bear \\ud83d\\udc3b | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $22B | $26B | $31B | $36B | $42B |\\n| Growth | 25.00% | 22.31% | 20.00% | 17.69% | 16.19% | 15% |\\n\\n| Base \\ud83d\\udcc8 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 |\\n| --- | --- | --- | --- | --- | --- | --- |\\n| Revenue | $18B | $22B | $27B | $33B | $40B | $48B |\\n| Growth | 25.00% | 23.66% | 22.50% | 21.34% | 20.60% | 20% | [...] Current revenue of $18B and 25% growth rate demonstrate Stripe's continued market strength, though growth has moderated from previous years' rapid expansion. This positions them well among B2B payment processors while reflecting a maturing business in the competitive fintech landscape.\\n\\n| Multiple | Valuation |\\n| --- | --- |\\n| 1x | $18B |\\n| 5x | $90B |\\n| 10x | $180B |\\n| 15x | $270B |\\n| 25x | $450B |\", \"score\": 0.80698997, \"raw_content\": null}, {\"url\": \"https://samboboev.medium.com/deep-dive-stripe-vs-adyen-comparing-2024-performance-ab585de2d27d\", \"title\": \"Deep Dive: Stripe vs. Adyen: Comparing 2024 Performance\", \"content\": \"In terms of revenue, Adyen\\u2019s business model focuses on \\u201cNet Revenue,\\u201d which excludes pass-through costs. Adyen\\u2019s net revenue for 2024 was \\u20ac1,996.1 million, a 23% YoY increase\\u200b. Stripe, as a private company, does not publicly break out revenue figures in its community letter; however, the company confirmed that it was profitable in 2024\\u200b. Stripe\\u2019s profitability in 2024 marked a notable milestone, enabling it to \\u201cplow back\\u201d operating earnings into R&D\\u200b. Adyen also demonstrated strong [...] Both Stripe and Adyen reported strong growth in 2024, with significant transaction volumes and improving profitability. Total Payment Volume (TPV) on Stripe\\u2019s platform reached $1.4 trillion for 2024, a 38% year-over-year (YoY) increase\\u200b. This makes Stripe responsible for about 1.3% of global GDP in payments processing\\u200b. Adyen\\u2019s processed volume was of a similar magnitude at \\u20ac1.29 trillion in 2024, which was up 33% YoY\\u200b. Adyen\\u2019s management notes that excluding a single large client (eBay), [...] This week, we\\u2019re diving into the battle of fintech giants: Stripe vs. Adyen. Both payment powerhouses had a stellar 2024, but they took different paths to success. Stripe saw $1.4 trillion in Total Payment Volume (TPV), growing 38% YoY, while Adyen wasn\\u2019t far behind with \\u20ac1.29 trillion processed (+33% YoY). Adyen maintained its 50% EBITDA margin, while Stripe finally hit full-year profitability, proving its business model can scale.\", \"score\": 0.77994305, \"raw_content\": null}, {\"url\": \"https://www.demandsage.com/stripe-statistics/\", \"title\": \"Stripe Revenue Statistics (2025): Market Share Data\", \"content\": \"Initially launched in the United States, Stripe has now expanded across 46 countries. However, the valuation of Stripe was reduced from $95 billion to $70 billion in the span of 3 years.\\n\\nInterested to know more details about Stripe for 2025?\\n\\nI have compiled the latest Stripe Statistics of 2025 in this article. So, let\\u2019s get into the details!\\n\\nStripe Statistics 2025: At A Glance\\n--------------------------------------- [...] According to John Collison (co-founder of Stripe), the company has over 40 businesses with a $1 billion payment volume each.\\n\\nFurther, according to the company data, there is two times increase in the payment volume of the Stripe enterprise clients.\\n\\nSource:Backlinko, Twitter, Stripe.\\n\\nHow Does Stripe Make Money?\\n------------------------------- [...] The stripe model charges a commission on every successful transaction. The standard rate of the charges made by Stripe is 2.9% of the transaction + 25 cents.\\n\\nHowever, the transaction fees or the commission rate may vary according to the plan the client has opted for.\\n\\nAdditionally, enterprise clients get a custom pricing option that tailors the features according to their requirements.\", \"score\": 0.76194656, \"raw_content\": null}, {\"url\": \"https://stripe.com/en-li/annual-updates/2024?__hsfp=1096049464&__hssc=106715356.1.1668124800362&__previewId&__s=1ngpkifa1w8cdaucpyke\", \"title\": \"Stripe's 2024 annual letter\", \"content\": \"Start-up incorporation\\n\\nCarbon removal\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Stripe\\u2019s 2024 annual letter\\n\\n2024 was a good year for the internet economy and for our users: in aggregate, their revenue processed on Stripe grew seven times faster than the revenue of companies in the S&P 500. [...] # Stripe logo\\n\\n# Global payments\\n\\nOnline payments\\n\\nFraud prevention\\n\\nAcceptance optimisations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation\\n\\nSales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n# More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStart-up incorporation\\n\\n# By stage\\n\\n# By business model [...] # By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\n# Get started\\n\\n# Guides\\n\\n# Global payments\\n\\nOnline payments\\n\\nFraud prevention\\n\\nAcceptance optimisations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation\\n\\nSales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\", \"score\": 0.7578844, \"raw_content\": null}, {\"url\": \"https://backlinko.com/stripe-users\", \"title\": \"Stripe Revenue and Growth Statistics (2024) - Backlinko\", \"content\": \"That\\u2019s all for my collection of Stripe stats for 2024. Stripe has quickly grown into one of the major players in the fintech industry, processing one trillion dollars annually.\\n\\nHopefully, you found these facts and figures interesting. As a private company, Stripe doesn\\u2019t disclose a lot of juicy data. But I did my best to collect what was out there.\\n\\nNext-level SEO training and link building strategies\\n\\n#### Learn\\n\\n#### Company\\n\\n#### Connect\\n\\n\\u00a9 2025 Backlinko is a Trademark of Semrush Inc [...] Here\\u2019s the breakdown of the number of available payment methods supported by Stripe in each category:\\n\\n| Category | Number of payment methods |\\n| --- | --- |\\n| Cards | 10 |\\n| Bank debits | 6 |\\n| Bank redirects | 10 |\\n| Bank transfers | 5 |\\n| Buy now, pay later | 5 |\\n| Real-time payments | 4 |\\n| Vouchers | 4 |\\n| Wallets | 15 |\\n\\nSource: Stripe\\n\\n## How Does Stripe Make Money?\\n\\nThe Stripe business model is built on charging a commission per successful transaction. [...] ## Stripe Enterprise Growth\\n\\nStripe has over 100 companies that process more than $1 billion via Stripe on an annual basis each.\\n\\nNumber of large customers processing more than $1 billion each has more than doubled from more than 40 firms reported earlier in 2020.\\n\\nSources: Stripe, X/Twitter\\n\\n## Who Founded Stripe?\\n\\nStripe was founded in 2010 by brothers Patrick and John Collison.\", \"score\": 0.7402687, \"raw_content\": null}, {\"url\": \"https://helplama.com/stripe-revenue-users-statistics/\", \"title\": \"Stripe Revenue and Users Statistics 2025\", \"content\": \"Stripe Revenue and Users Statistics 2025\\n========================================\\n\\nImage 2: Stripe Revenue Users Stats\\n\\nLast Updated:August 2025\\n\\nStripe, the leading online payments platform, has been on an incredible trajectory, revolutionizing the way businesses transact in the digital age. As we dive into the year 2025, it\\u2019s time to uncover the astonishing Stripe statistics that highlight its unstoppable rise in revenue and user engagement. [...] Stripe Revenue 2024\\n-----------------------\\n\\n### Stripe Annual Revenue\\n\\nStripe revenue increased by 20% to $14 billion in 2022 from $12 billion in 2021.\\n\\nSpeaking of Stripe\\u2019s revenue growth, the payment giant\\u2019s gross revenue has seen a compound annual growth rate (CAGR) of 69.52% since 2017.\\n\\nFinancial yearRevenue\\n2014$40 million\\n2016$450 million\\n2017$1 billion\\n2018$1.5 billion\\n2019$2 billion\\n2020$7.4 billion\\n2021$12 billion\\n2022$14 billion\\n\\nSource: Backlinko, Sacra.com\\n\\nImage 3: Stripe Growth [...] When current customers are combined with historical ones, the number is estimated at 3,820,274.\\n\\nSource: Builtwith.com\\n\\n### What Companies Use Stripe?\\n\\nOver 100 companies use Stripe to process over $1 billion annually.\\n\\nCompany NameWebsiteRevenueLocation\\nFlexport www.flexport.com$1 Billion United States\\nInstacart www.instacart.com$1 Billion United States\\nShopify www.shopify.com$4 Billion Canada\\nLyft www.lyft.com$2 Billion United States\\nWayfair www.wayfair.com$14 Billion United States\", \"score\": 0.72465646, \"raw_content\": null}, {\"url\": \"https://www.forbes.com/companies/stripe/\", \"title\": \"Stripe | Company Overview & News - Forbes\", \"content\": \"Business Services & Supplies\\n\\n#183\\n\\nAmerica's Top Private Companies (2024)\\n\\nBusiness Services & Supplies\\n\\n#3\\n\\nCloud 100 (2024)\\n\\nFinancial Infrastructure\\n\\nMore Lists\\n\\n### Stripe Financial Summary\\n\\nSelect list year\\n\\n2024\\n\\n2023\\n\\n2022\\n\\nRevenue\\n\\n$3.5B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRevenue\\n\\n$3.1B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRevenue\\n\\n$2.5B\\n\\nSources\\n\\nFactSet; Forbes.\\n\\nRelated People & Companies\\n\\nImage 1: John Collison\\n\\nJohn Collison\\n\\nEmployed by Company\\n\\nView Profile\\n\\nImage 2: Peregrine\\n\\nPeregrine [...] Founded in 2009 by Irish brothers Patrick and John Collison, Stripe processes payments for online businesses. In 2024, its subscription-billing product reached more than 300,000 business customers, and it agreed to acquire Bridge, which makes infrastructure software for cryptocurrency-based stable coins (tokens that are typically pegged to the value of the U.S. dollar), for $1.1 billion. Stripe counts Nvidia, OpenAI and Pepsi among its customers. [...] Stripe | Company Overview & News\\n\\n===============\\n\\n, for $1.1 billion. Stripe counts Nvidia, OpenAI and Pepsi among its customers. \\nFunding: $2.2 billion from Sequoia, General Catalyst, Founders Fund, Thrive and others.\\n\\nLatest valuation: $70 billion, based on a September 2024 sha... Read More\", \"score\": 0.67192334, \"raw_content\": null}, {\"url\": \"https://en.wikipedia.org/wiki/Stripe,_Inc.\", \"title\": \"Stripe, Inc. - Wikipedia\", \"content\": \"On 4 February 2025, Stripe announced it had acquired Bridge.( The acquisition of the two-year-old stablecoin platform company is valued at $1.1 billion.(\\n\\nInvestments\\n-----------\\n\\n[edit] [...] On 26 July 2024, Stripe acquired a Merchant of Record services start-up Lemon Squeezy&action=edit&redlink=1 \\\"Lemon Squeezy (startup) (page does not exist)\\\").(\\n\\nSee also\\n--------\\n\\n[edit]\\n\\n   Electronic commerce\\n   List of online payment service providers\\n   Payment gateway\\n   Payment service provider\\n   Subscription business model\\n\\nReferences\\n----------\\n\\n[edit] [...] 80.   ^\\\"Stripe completes Bridge acquisition\\\". _Stripe, Inc_. 4 February 2025. Retrieved 8 April 2025.\\n81.   ^Sigalos, MacKenzie (4 February 2025). \\\"Stripe closes $1.1 billion Bridge deal, prepares for aggressive stablecoin push\\\". _CNBC_. Retrieved 8 April 2025. Stripe closed its $1.1 billion purchase of Bridge Network.\", \"score\": 0.5923135, \"raw_content\": null}, {\"url\": \"https://stripe.com/\", \"title\": \"Stripe | Financial Infrastructure to Grow Your Revenue\", \"content\": \"- Jobs\\n- Newsroom\\n\\n- Stripe Press\\n- Become a partner\\n\\nSessions 2025\\n\\u00b7\\nWatch on demand\\n\\n# Financial infrastructure to grow your revenue\\n\\nFinancial infrastructure to grow your revenue\\n\\nFinancial infrastructure to grow your revenue\\n\\nJoin the millions of companies that use Stripe to accept payments online and in person, embed financial services, power custom revenue models, and build a more profitable business.\\n\\nStart now\\nContact sales\\n\\nStart now\\nContact sales\\n\\nRocket Rides\\n\\nSearch\\n\\nToday [...] # The backbone for global commerce\\n\\nStripe makes moving money as easy and programmable as moving data.\\nOur teams are based in offices around the world and we process hundreds of billions of dollars each year for ambitious businesses of all sizes.\\n\\n# 500M+\\n\\nAPI requests per day, peaking at 13,000 requests a second.\\n\\n# 99.999%\\n\\nhistorical uptime for Stripe services.\\n\\n# 90%\\n\\nof U.S. adults have bought from businesses using Stripe.\\n\\n# 135+\\n\\ncurrencies and payment methods supported.\", \"score\": 0.5360463, \"raw_content\": null}, {\"url\": \"https://onstrategy.eu/stripes-2025-letter-explained/\", \"title\": \"Stripe's 2025 Letter explained - onStrategy\", \"content\": \"Stripe\\u2019s acquisition of Bridge, the leading stablecoin orchestration platform, signals a massive shift in financial infrastructure. Stablecoins make money movement cheaper, faster, and programmable, and Stripe is betting that they will become a core part of business transactions worldwide. While stablecoins have been largely a crypto-native phenomenon, Stripe\\u2019s push into the space suggests that they\\u2019re about to become mainstream for corporate treasury, global payroll, and real-world commerce, [...] , Traxero (for tow truck operators), and Clio (for law firms), all examples of vertical SaaS businesses that act as financial platforms for their industries. This is a fundamental shift: Stripe it\\u2019s enabling software-driven economies at every level, including small businesses that previously couldn\\u2019t access this kind of technology.\\n\\n7/ Stablecoins are Stripe\\u2019s next big bet, because money itself is changing.\", \"score\": 0.5106191, \"raw_content\": null}], \"response_time\": 9.53, \"request_id\": \"adf9499f-4b2e-4a4d-8ca5-1ae55f579f20\"}, \"error\": null}}}",
      "message_id": "e5e48ad5-fa67-4982-83f7-1bcbe9aabe5e"
    },
    {
      "role": "assistant",
      "content": "<function_calls>\n<invoke name=\"web_search\">\n<parameter name=\"query\">Stripe recent news announcements 2024 2025 acquisitions partnerships</parameter>\n<parameter name=\"num_results\">10</parameter>\n</invoke>\n</function_calls>",
      "tool_calls": null,
      "message_id": "c6a9b015-c903-4ad3-b768-17b665130f6b"
    },
    {
      "role": "user",
      "content": "{\"tool_execution\": {\"function_name\": \"web_search\", \"xml_tag_name\": \"web-search\", \"tool_call_id\": null, \"result\": {\"success\": true, \"output\": {\"query\": \"Stripe recent news announcements 2024 2025 acquisitions partnerships\", \"follow_up_questions\": null, \"answer\": \"Stripe has been highly active with acquisitions and partnerships throughout 2024 and 2025, completing several major deals that expand its capabilities across different payment sectors. The company acquired Bridge, a stablecoin platform, in February 2025 in what represents Stripe's largest acquisition to date, positioning the company to leverage the $15.6 trillion stablecoin market that moved in 2024. Other significant acquisitions include Orum, a payments orchestration platform for instant payment capabilities, acquired in July 2025, and Lemon Squeezy, which was acquired in 2024 to enhance digital product sales capabilities. Stripe also announced plans to acquire Recko to automate payments reconciliation and OpenChannel to expand its platform offerings. On the partnership front, Stripe has deepened its collaboration with NVIDIA to enhance AI-powered capabilities and expand global access to NVIDIA's AI platform, while also establishing partnerships with Klarna to help online businesses grow revenue, Salesforce for global commerce, and Shopify to enable stablecoin payments for millions of merchants. The company processed over $1.4 trillion in total payment volume in 2024, representing 38% growth, and now serves half of the Fortune 100 companies, demonstrating its evolution from a startup-focused platform to a major enterprise player.\", \"images\": [\"https://images.stripeassets.com/fzn2n1nzq965/1ZK0vLagjDOXEdwyApsNIQ/22fd18c98bcb8ca7d550ebf6c08678d4/Newsroom_carousel-bridge_1080x736.jpg?q=80&w=1080&fm=jpg\", \"https://7834608.fs1.hubspotusercontent-na1.net/hub/7834608/hubfs/Slide4-May-13-2024-11-41-48-2955-AM.png?width=1200&length=1200&name=Slide4-May-13-2024-11-41-48-2955-AM.png\", \"https://images.stripeassets.com/fzn2n1nzq965/kxTRUol9vTSEnrwOQAbpE/98effe9be044e01843344712996333f4/Clear___Stripe_4000x2000_v2.png?q=80&w=552&fm=jpg\", \"https://media.licdn.com/dms/image/v2/D4D10AQHe4_8XoctNJw/image-shrink_1280/image-shrink_1280/0/1732536005800?e=2147483647&v=beta&t=W7jPeXOVfHtljCWml648GKWc2y8_5loZjNui1pbsocg\", \"https://images.stripeassets.com/fzn2n1nzq965/77hV1QeIA0vzLaAzhgtPtQ/458eca960412cd1b4fd7d519a0f98f12/NRP_NVIDIA.png?q=80&w=1080&fm=jpg\"], \"results\": [{\"url\": \"https://www.fintechfutures.com/m-a/stripe-acquires-orum\", \"title\": \"Stripe acquires paytech start-up Orum - FinTech Futures\", \"content\": \"Orum is a payments orchestration platform that connects businesses to instant payment capabilities via API.\\n\\nPicture of Tyler Pathe\\n\\nJuly 24, 2025\\n\\nOrum - FinTech News\\n\\nOrum, a payments orchestration platform connecting businesses to instant payment capabilities via API, is being acquired by US fintech giant Stripe for an undisclosed sum. [...] .png?width=300&auto=webp&quality=80&disable=upscale)\\nFinTech Futures\\n\\n##### Events\\n\\n##### Resources\\n\\n##### Working With Us\\n\\n##### Join Us\\n\\n##### Follow Us\\n\\nInforma\\n\\nCopyright \\u00a9 2025 Informa PLC. Informa PLC is registered in England and Wales with company number 8860726 whose registered and head office is 5 Howick Place, London, SW1P 1WG. [...] In a blog post announcing the acquisition, Kirkpatrick states: \\\"After thoughtful consideration, it became clear that we have a rare opportunity to accelerate Orum\\u2019s mission and greatly increase our impact by becoming part of Stripe, which processed more than $1.4 trillion in total payment volume in 2024.\\\"\", \"score\": 0.776708, \"raw_content\": null}, {\"url\": \"https://www.chargeflow.io/blog/stripe-statistics\", \"title\": \"Stripe Stats for 2025: The Most Accurate Stripe Statistics Online\", \"content\": \"### Full List of Stripe Acquisitions to Date\\n\\nStripe has acquired 20 SaaS, Enterprise Software, and Enterprise Tech companies in the \\u00c2 U.S. and other countries. In classic Stripe fashion, most of the investment amounts are undisclosed. Here are 19 Stripe acquisitions that have currently reached our radar:\\n\\n### 20 Hottest Stripe Investments to Date [...] Stripe\\u00e2\\u0080\\u0099s BFCM stats are always a thing for the books and 2024 was no different. Businesses processed more than $31 billion on Stripe from Black Friday through Cyber Monday, the company proclaimed. That\\u00e2\\u0080\\u0099s a notable uptick from $18.6 billion in 2023.\\n\\n## Stripe Investment & Acquisition Stats\\n\\nStripe has carried out 18 acquisitions since 2013. The payment giant has also funded several tech companies since 2017. Some of these strategic investments reach hundreds of millions of U.S. dollars. [...] Stripe continues to fire on all cylinders after achieving unicorn status in just three years. The payment giant is one of the fastest-growing fintechs worldwide, with customers including Apple, Alaska Airlines, Best Buy, Lotus Cars, Microsoft, Uber, and Zara.\\n\\n## Stripe Revenue Statistics\", \"score\": 0.7371019, \"raw_content\": null}, {\"url\": \"https://a16z.com/newsletter/what-stripes-acquisition-of-bridge-means-for-fintech-and-stablecoins-april-2025-fintech-newsletter/\", \"title\": \"What Stripe's Acquisition of Bridge Means for Fintech and Stablecoins\", \"content\": \"###### TOP SUGGESTIONS\\n\\n# What Stripe\\u2019s Acquisition of Bridge Means for Fintech and Stablecoins (April 2025 Fintech Newsletter)\\n\\nJames da Costa and Sam Broner\\n\\n#### What Stripe\\u2019s Acquisition of Bridge Means for Fintech and Stablecoins [...] their merchants a cheaper alternative to credit card payments. In 2024, Stripe grew 38% \\u2014 surpassing $1.4 trillion in total paymentsvolume; the Bridge acquisition could therefore fuel further global growth through stablecoins. [...] In February, Stripe completed its acquisition of stablecoin platform Bridge \\u2014 its largest acquisition to date \\u2014 following the reintroduction of crypto payments on Stripe for U.S. businesses last year. More broadly, after years of steady stablecoin growth, this acquisition represents the payment industry\\u2019s first big acknowledgement that stablecoins can go further mainstream. In 2024, stablecoins moved $15.6 trillion in value \\u2014 putting stablecoin transaction volume on par with Visa\\u2019s.\", \"score\": 0.72395444, \"raw_content\": null}, {\"url\": \"https://stripe.com/newsroom/news\", \"title\": \"Stripe Newsroom: The Latest News & Announcements\", \"content\": \"# Stripe Capital drives sharp increases in small businesses\\u2019 revenue growth\\n\\n## Corporate\\n\\n# Stripe acquires OpenChannel\\n\\n## Product\\n\\n# Stripe expands Stripe Terminal across Europe\\n\\n## Corporate\\n\\n# Matt Huang joins Stripe\\u2019s board of directors\\n\\n## Product\\n\\n# Klarna partners with Stripe to help online businesses grow their revenue\\n\\n## Corporate\\n\\n# Stripe will acquire Recko to help internet businesses automate payments reconciliation\\n\\n## Product [...] # Stripe announces expanded interoperability, major upgrades to Stripe Billing, new partnerships with NVIDIA and Pepsi, and more\\n\\n## Corporate\\n\\n# Stripe deepens collaboration with NVIDIA to enhance Stripe\\u2019s AI-powered capabilities and expand global access to NVIDIA\\u2019s AI platform\\n\\n## Product\\n\\n# Stripe expands in Toronto to accelerate Canadian business growth\\n\\n## Product\\n\\n# Enercare chooses Stripe as its strategic payments partner\\n\\n## Product [...] ## Corporate\\n\\n# Stripe launches world\\u2019s first large-scale carbon removal purchase tool\\n\\n## Corporate\\n\\n# Stripe will acquire Paystack to accelerate online commerce across Africa\\n\\n## Corporate\\n\\n# Stripe named a Leader by Forrester, strikes global commerce partnership with Salesforce\\n\\n## Corporate\\n\\n# Dhivya Suryadevara to join Stripe as Chief Financial Officer\\n\\n## Corporate\\n\\n# Stripe hires Mike Clayville as Chief Revenue Officer\\n\\n## Product\", \"score\": 0.7140107, \"raw_content\": null}, {\"url\": \"https://stripe.com/blog/top-product-updates-sessions-2025\", \"title\": \"Our top product updates from Sessions 2025\", \"content\": \"Stripe Profiles\\n\\nIn addition to sharing updates from our four core product areas, we also announced several other features, products, and programs that help you grow. We announced a new API that lets you embed agentic commerce directly into your applications, unveiled Stripe Verified, a new credential for an accredited Stripe business, and previewed a new networking capability that lets you easily transact with other businesses on Stripe. Here\\u2019s what\\u2019s new: [...] More than 300,000 businesses use Stripe Billing\\u2014the heart of our Revenue suite\\u2014including the world\\u2019s leading AI companies such as OpenAI, Anthropic, Cursor, Midjourney, and now, NVIDIA. We unveiled the biggest updates to our Revenue suite yet, including two new extensibility primitives\\u2014Scripts and Workflows\\u2014that turn Stripe into a programmable revenue engine, upgrades to usage-based billing to help you better monetize your AI offerings, and the evolution of Stripe Tax into an all-in-one global [...] This morning at our annual conference, Stripe Sessions, we shared the biggest changes we\\u2019ve ever made to Stripe with thousands of business leaders and builders. Our keynote covered our progress in four core product areas: Payments, Connect, Revenue, and Money Management. We highlighted how we\\u2019re helping you grow by applying frontier technologies like AI and stablecoins to your commerce stacks, by leveraging the vast scale of the Stripe network, which now comprises millions of businesses\", \"score\": 0.68191797, \"raw_content\": null}, {\"url\": \"https://www.lemonsqueezy.com/blog/stripe-lemon-squeezy-update-2025\", \"title\": \"Stripe + Lemon Squeezy Update: A Big Milestone Reached\", \"content\": \"### Lemon Squeezy Black Friday & Cyber Monday Roundup 2024 \\u00e2\\u009c\\u00a8\\n\\n### Resolved Security Vulnerability\\n\\n### Juicy news! Payout fees sliced\\n\\n### \\u00f0\\u009f\\u008d\\u008b Stripe acquires Lemon Squeezy\\n\\n### Tailored Pricing for Every User: Unveiling Consumption Billing at LemonSqueezy\\n\\n### Lemon Drop (August 2024)\\n\\n### John O'Nolan \\u00e2\\u0080\\u0094 Make Lemonade Podcast\\n\\n### Using AI to improve Lemon Squeezy's Customer Support\\n\\n### Introducing PayPal Subscriptions [...] When Stripe acquired Lemon Squeezy last year, we knew it was the start of something special. We\\u00e2\\u0080\\u0099ve always shared a deep love for our customers and a mission to make selling digital products simple, elegant, and joyful. Not too tangy, not too sweet, but just right. Since then, we\\u00e2\\u0080\\u0099ve been blending the best of both worlds and now, we\\u00e2\\u0080\\u0099re excited to share the first major milestone in that journey.\\u00e2\\u0080\\u008d [...] When Stripe acquired Lemon Squeezy last year, we knew it was the start of something special. We\\u00e2\\u0080\\u0099ve always shared a deep love for our customers and a mission to make selling digital products simple, elegant, and joyful. Not too tangy, not too sweet, but just right.\\n\\nSince then, we\\u00e2\\u0080\\u0099ve been blending the best of both worlds and now, we\\u00e2\\u0080\\u0099re excited to share the first major milestone in that journey.\\n\\n## Introducing Stripe Managed Payments\\u00c2\", \"score\": 0.6191726, \"raw_content\": null}, {\"url\": \"https://techcrunch.com/2025/02/27/stripe-finalizes-tender-sale-at-a-91-5b-valuation-says-payment-volumes-grew-to-1-4t-in-2024/\", \"title\": \"Stripe finalizes tender sale at a $91.5B valuation, says payment ...\", \"content\": \"The news coincides with Stripe\\u2019s annual letter penned by CEO and co-founder Patrick Collison, which noted that payment volume in 2024 grew to $1.4 trillion, up 38% on the year before.\\n\\nIt\\u2019s a big number for Stripe, but to put it into some context, Visa said that its payment volume for 2024 was $13.2 trillion. The margin that Stripe makes on transactions (which are what make up payment volume) remains thin, and so this business remains one where it still has room to (and needs to) scale. [...] Stripe also added that it is now used by half of the Fortune 100 companies, underscoring how it has catapulted from a startup working with other startups, into a major enterprise player.\\n\\nThe tender offer was originally reported to be in the works earlier this month.\\n\\n### Tech and VC heavyweights join the Disrupt 2025 agenda [...] ### Security flaws in a carmaker\\u2019s web portal let one hacker remotely unlock cars from anywhere\\n\\nSecurity flaws in a carmaker\\u2019s web portal let one hacker remotely unlock cars from anywhere\\n\\n### The hidden cost of living in Mark Zuckerberg\\u2019s $110M compound\\n\\nThe hidden cost of living in Mark Zuckerberg\\u2019s $110M compound\\n\\n### The computer science dream has become a nightmare\\n\\nThe computer science dream has become a nightmare\\n\\nTechCrunch Logo\\n\\n\\u00a9 2025 TechCrunch Media LLC.\", \"score\": 0.48177138, \"raw_content\": null}, {\"url\": \"https://stripe.com/blog/biggest-updates-sessions-2024\", \"title\": \"Our biggest product updates from Sessions 2024 - Stripe\", \"content\": \"This morning at our annual conference, Stripe Sessions, we shared our biggest product updates of the year with thousands of business leaders and builders who gathered at the Moscone Center in San Francisco. Our keynote covered Stripe\\u2019s progress in our three core product areas and shared how we\\u2019re evolving Stripe to interoperate with other payments providers. [...] This post recaps the top announcements and product releases. For a comprehensive list of everything that\\u2019s new since Sessions last year, see our changelog.\\n\\n## Increase conversion and reduce costs with Global Payments [...] Sales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStartup incorporation\\n\\nCarbon removal\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Our biggest product updates from Sessions 2024\\n\\nBlog > Sessions 2024 > Header image\", \"score\": 0.4654124, \"raw_content\": null}, {\"url\": \"https://stripe.com/sessions/2024\", \"title\": \"Sessions 2024 - Stripe\", \"content\": \"Even the briefest system outages can have unforeseen orders of impact for businesses. Learn best practices from industry leaders who manage reliability at scale, and go inside the culture and systems at Stripe that maintain reliability while processing 1% of the global GDP.\\n\\n# Registration for Sessions 2026 is now open\\n\\nJoin us April 28\\u201330 in San Francisco to help shape the future of the internet economy.\\n\\nSessions 2025 > Registration banner > Image\\n\\n# Looking for more information? Contact us. [...] Sales tax & VAT automation\\n\\nOnline invoices\\n\\nCustom reports\\n\\nData sync\\n\\n## More\\n\\nAccess to 100+ globally\\n\\nAccelerated checkout\\n\\nLinked financial account data\\n\\nOnline identity verification\\n\\nStartup incorporation\\n\\nCarbon removal\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\nStart integrating Stripe\\u2019s products and tools\\n\\n# Get started\\n\\n# Guides\\n\\n# Sessions 2024\\n\\nKeynotes\\n\\n## Opening keynote\\n\\nOpening keynote [...] People like beautiful things for rational reasons. See why well-crafted products are expressions of care and dedication\\u2014and how that correlates to business success.\\n\\nTrends and inspiration\\n\\n## Building a culture of system reliability\\n\\nBuilding a culture of system reliability\", \"score\": 0.41905406, \"raw_content\": null}, {\"url\": \"https://stripe.com/newsroom\", \"title\": \"Stripe Newsroom: The Latest News & Announcements\", \"content\": \"# Stripe Radar now protects ACH and SEPA payments from fraud\\n\\n# Stories\\n\\nBusinesses around the world are using Stripe.\\n\\n# Stripe and the UK: A decade in the digital economy\\n\\n# An interview with the cofounder of PhotoRoom\\n\\nMatthieu Rouif talks about the practical applications of AI magic in commercial and resale photography\\u2014and how to maintain focus in a surging field.\\n\\n# Mr & Mrs Smith promises the perfect stay for hotel lovers\\n\\n# Trustap lets strangers transact with confidence [...] Online identity verification\\n\\nStartup incorporation\\n\\nFinancial accounts\\n\\n# By stage\\n\\n# By business model\\n\\n# By use case\\n\\n# By industry\\n\\n# Ecosystem\\n\\n# Get started\\n\\n# Guides\\n\\n# Global payments\\n\\nOnline payments\\n\\nIn-person payments\\n\\nFraud prevention\\n\\nAcceptance optimizations\\n\\n# Money Management\\n\\nPayments for platforms\\n\\nSend money to third parties\\n\\nCustomer financing\\n\\nPhysical and virtual cards\\n\\n# Revenue and Finance Automation\\n\\nSubscriptions and usage-based\\n\\nAccounting automation [...] # Follow Stripe\\n\\n# News\\n\\n## Product\\n\\n# Stripe launches new products to accelerate growth for businesses in Germany\\n\\n## Product\\n\\n# Stripe rolls out 15 new features to boost France\\u2019s digital economy\\n\\n## Product\\n\\n# Stripe will help millions of Shopify merchants to accept stablecoin payments\\n\\n## Product\\n\\n# Stripe fuels UK business growth with global money management capabilities\\n\\n## Product\\n\\n# Stripe accelerates the utility of AI and stablecoins with major launches\\n\\n## Product\", \"score\": 0.37917203, \"raw_content\": null}], \"response_time\": 10.1, \"request_id\": \"05e0e7f5-9eca-4b1b-ab06-a385dbf597e0\"}, \"error\": null}}}",
      "message_id": "7b4bd5b4-d617-4c10-be51-a3cc34e2fd88"
    }
  ],
  "input_token_count": 58080,
  "compressed_token_count": 54826,
  "compression_ratio": 0.55
}