fix(board_worker): goal-task DoD 'zero NEW failures' not 'zero total'#419
Merged
Conversation
… — unblock self-verify A goal task did clean, mergeable work but failed its own verification: the team_executor verifier's LLM-derived criterion was 'Zero test failures or skipped tests' against the FULL suite, which has 5 pre-existing failures + 21 skips (tests/ root + sandbox-gated tests CI doesn't run). The agent flagged 'zero NEW failures, branch clean, ready for PR' but rejected itself on the strict bar. Reword _append_definition_of_done criteria 3+4: your change must introduce ZERO NEW failures; pre-existing/unrelated failures+skips are OUT OF SCOPE (don't block on or fix them); the merge gate is the repo's REQUIRED CI checks, not a fully-green pre-existing local suite. The stage_planner (which derives acceptance_criteria from the goal text) now generates a no-regression criterion the verifier can satisfy. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
A goal task (89fdd864) did clean, mergeable work but failed its own self-verification: team_executor's verifier checked the LLM-derived acceptance criterion 'Zero test failures or skipped tests' against the full suite, which has 5 pre-existing failures + 21 skips (tests/ root + sandbox-gated tests CI doesn't run). The agent correctly noted 'zero NEW failures, branch is clean, ready for PR' but then rejected itself on the over-strict bar — an autonomy stall from mis-specified done-ness, not bad work.
Root cause:
_append_definition_of_donesaid 'run the test suite and make them pass / verified green'; team_executor'sstage_planner(it LLM-derives acceptance_criteria from the goal text — not hardcoded) turned that into 'zero failures'.Fix
Reword DoD criteria 3+4: your change must introduce ZERO NEW failures; pre-existing/unrelated failures+skips are OUT OF SCOPE (don't block on or fix them); the merge gate is the repo's REQUIRED CI checks, not a fully-green pre-existing local suite. The planner now generates a no-regression criterion the verifier can actually satisfy. Test asserts the new intent.
Note: the 5 pre-existing full-suite failures are a separate repo-health item (CI runs only
tests/unit, which is green).🤖 Generated with Claude Code