fix: clarify judge prompt to not penalize agent tool use#375
Merged
olearycrew merged 1 commit intomainfrom May 4, 2026
Merged
Conversation
The judge prompt's 'Do NOT use any tools' instruction was ambiguous — some judge models interpreted it as a constraint on the evaluated agent and auto-failed tasks when they saw tool calls in transcripts. Scoped the 'no tools' rule explicitly to the grader and added a clarification that agent tool usage is normal and expected. Closes #374
Contributor
Code Review SummaryStatus: No Issues Found | Recommendation: Merge The change is well-targeted and the fix is clear. Renaming the header and adding the Files Reviewed (1 file)
Reviewed by claude-4.6-sonnet-20260217 · 70,797 tokens |
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #374
Problem
The judge prompt includes 'Do NOT use any tools' as an instruction to the grader LLM, but some models interpret this as a constraint on the evaluated agent and auto-fail tasks when they see tool calls in transcripts.
Fix
Two changes to
_build_judge_prompt()inscripts/lib_grading.py:Renamed the header from
CRITICAL RULES→CRITICAL RULES FOR YOU, THE GRADER (not the agent being graded)to scope the no-tools instruction explicitly.Added an
IMPORTANTparagraph clarifying that the agent being graded may have used tools during execution and that this is normal — the no-tools rules apply only to the grader itself.🤖 This PR was opened by @olearycrew's OpenClaw bot. Please review carefully!