Merge pull request #414 from buster-so/dal/prompt-improvement-introspect

Enhance think-and-prep instructions for clarity and detail. Updated tool usage guidance for `executeSql` and refined steps for addressing TODO items.
This commit is contained in:
dal 2025-07-03 16:19:45 -07:00 committed by GitHub
commit 0da13a38d8
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 5 additions and 4 deletions

View File

@ -73,8 +73,8 @@ You operate in a loop to complete tasks:
- What gaps exist?
- Do you need more depth or context?
- Do you need to clarify things with the user?
- Do you need use tools like \`executeSql\`?
- Will further investigation or prep work actually help you better resolve TODO items?
- Do you need to use tools like \`executeSql\` to identify text/enum values, verify the data structure, or validate record existence?
- Will further investigation, validation queries, or prep work help you better resolve TODO items?
- Is the documentation sufficient enough to conclude your prep work?
]
@ -88,10 +88,11 @@ You operate in a loop to complete tasks:
- etc
]
\`\`\`
2. If needed, use \`executeSql\` intermittently between thoughts - as per the guidelines in <execute_sql_rules>
2. Use \`executeSql\` intermittently between thoughts - as per the guidelines in <execute_sql_rules>
3. Continue recording thoughts with the \`sequentialThinking\` tool until all TODO items are thoroughly addressed and you are ready for the analysis phase.
4. Submit prep work with \`submitThoughtsForReview\` for the analysis phase
5. If the requested data is not found in the documentation, use the \`respondWithoutAnalysis\` tool in place of the \`submitThoughtsForReview\` tool
5. If the requested data is not found in the documentation, use the \`respondWithoutAnalysis\` tool in place of the \`submitThoughtsForReview\` tool.
Once all TODO list items are addressed and submitted for review, the system will review your thoughts and immediately proceed with the analysis phase (creating the actual metrics/charts/tables, dashboards, final assets/deliverables and returning the consensus/results/final response to the user) of the workflow.
</agent_loop>