fix: Enhance LLM tip generation robustness #56
Merged
gjt-prog merged 5 commits intoAgentToolkit:mainfrom Feb 19, 2026
Merged
Conversation
…t messages and handling malformed/empty responses, and add validation for trajectory data.
|
No actionable comments were generated in the recent review. 🎉 📝 WalkthroughWalkthroughAdded a module logger and improved error handling in Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~12 minutes Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Fix all issues with AI agents
In `@kaizen/llm/tips/tips.py`:
- Around line 152-156: Wrap the call to
TipGenerationResponse.model_validate(json.loads(clean_response)) in an
additional except block that catches pydantic.ValidationError (import
ValidationError from pydantic alongside other imports), log a warning similar to
the JSONDecodeError case using logger.warning with the error and a truncated
repr(clean_response[:500]), and return [] so the function gracefully handles
schema validation failures; ensure the new except references
TipGenerationResponse.model_validate, logger.warning, and clean_response.
🧹 Nitpick comments (2)
kaizen/llm/tips/tips.py (2)
78-80: Consider logging when skipping non-empty assistant content.The previous code raised an exception for unrecognized content, which served as an alert for unexpected data shapes. Silently continuing is fine for empty/
Nonecontent, but ifcontentis a non-empty, non-string, non-list value (e.g., a dict or unexpected type), this will now silently discard it. A debug-level log would help surface unexpected content shapes without breaking the flow.💡 Suggested improvement
else: - # Skip empty assistant messages (common from tool-calling patterns) - continue + # Skip empty assistant messages (common from tool-calling patterns) + if content: + logger.debug(f"Skipping unexpected assistant content type: {type(content).__name__}") + continue
5-5: Logger name"entities-mcp"looks like it belongs to a different module.The conventional Python pattern is
logging.getLogger(__name__), which would yield"kaizen.llm.tips.tips"and integrate naturally with hierarchical logger configuration. Using a hardcoded name from an unrelated namespace makes filtering and log routing confusing.💡 Suggested fix
-logger = logging.getLogger("entities-mcp") +logger = logging.getLogger(__name__)
…d validate trajectory data input.
visahak
previously approved these changes
Feb 16, 2026
Collaborator
|
LGTM. |
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@kaizen/llm/tips/tips.py`:
- Line 16: The module-level logger is created with a hardcoded name
"entities-mcp" which is incorrect; update the logger instantiation in tips.py
(the logger variable) to use the module's __name__ (logging.getLogger(__name__))
so logs inherit the package logger hierarchy and are attributable to
kaizen.llm.tips.tips instead of a copy-pasted name.
|
Caution Failed to replace (edit) comment. This is likely due to insufficient permissions or the comment being deleted. Error details |
visahak
previously approved these changes
Feb 17, 2026
vinodmut
approved these changes
Feb 17, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Enhance LLM tip generation robustness by skipping empty assistant messages and handling malformed/empty responses, and add validation for trajectory data.
Summary by CodeRabbit
Bug Fixes
Chores