feat(prompts): Enhance Task tool usage guidance in system prompts #761
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Add explicit instructions to proactively use the TASK tool for large, complex tasks that might fill the context window - Update Task Management section with specific guidance on when to use the TASK tool - Enhance Software Engineering Tasks workflow to encourage proactive TASK tool usage - Strengthen Subagent Delegation guidance for context overflow prevention - Update getSubagentSystemReminder to emphasize critical importance of using TASK tool for large tasks - Enhance Task tool description to emphasize breaking down large tasks that might fill the context window - Update usage notes to encourage proactive use for large, complex tasks and context overflow prevention This addresses the need to encourage the agent to break down large tasks into manageable chunks using the TASK tool, preventing context window overflow. Fixes #000 Co-authored-by: Qwen-Coder <[email protected]> The changes refine the usage of the Task tool to encourage agents to break down large tasks into manageable chunks, preventing context window overflow. The updates enhance both system prompts and tool descriptions to emphasize proactive use of the TASK tool for complex tasks that might overwhelm the agent's context window.
Co-authored-by: Qwen-Coder <[email protected]>
The CI was failing on Windows due to inconsistent line ending handling in integration-tests/test-helper.ts. The code was splitting content using the platform-specific EOL constant but then joining content back using hardcoded \n instead of the platform-appropriate EOL. This fix ensures consistent line ending handling across platforms by using the EOL constant for both splitting and joining operations. Co-authored-by: Qwen-Coder <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
TLDR
Enhance the usage guidance for the TASK tool in system prompts to encourage agents to proactively break down large, complex tasks into manageable chunks, preventing context window overflow.
Dive Deeper
This PR updates the system prompts and Task tool descriptions to make the agent more proactive about using the TASK tool for large, complex tasks. The changes include:
Reviewer Test Plan
npm run test:ci)Testing Matrix
(All testing indicators should be marked as validated after confirming the changes work properly)
Linked issues / bugs
This PR addresses the need to encourage the agent to break down large tasks into manageable chunks using the TASK tool, preventing context window overflow.
The changes have been implemented, tested (with snapshots updated), and committed to the branch. You would now use the GitHub web interface to create the pull request from the task-tool-usage-improvements branch to the main branch.