v1.6.0: make the gate actually gate — FAIL-level must-fix checks, equation numbering, delivery gate - #4
Merged
Merged
Conversation
…ation numbering, finalize_docx.py Third real-world failure: Codex delivered gridded tables, italic formula digits, and unnumbered/unreferenced equations — and the audit 'passed'. Root causes fixed: - WARN never blocked delivery: every Must-Fix rule promoted to FAIL (TABLE_BORDERS, TABLE_RULES, FORMULA_DIGIT_ITALIC, FORMULA_MULTILETTER_ITALIC, EQUATION_NUMBER_CENTER/TABS, CAPTION_ALIGN) - TABLE_BORDERS ignored style-driven grids (Table Grid lives in styles.xml); now resolves the style chain - No check existed for missing equation numbers or missing in-text citations: new EQUATION_UNNUMBERED (FAIL) + EQUATION_NOT_REFERENCED (WARN) - Rules in prose are not executed: scripts/finalize_docx.py is the one mandatory command (mechanical fixes + audit + DELIVERY GATE verdict, non-zero exit on FAIL); SKILL.md and agents/openai.yaml wired to it Verified against a reproduction of the Codex failures: all three now FAIL and the gate blocks delivery. 136 tests pass. Co-Authored-By: Claude Fable 5 <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes why the latest Codex run shipped gridded tables, italic formula digits, and unnumbered equations while the audit said pass:
TABLE_BORDERS,TABLE_RULES,FORMULA_DIGIT_ITALIC,FORMULA_MULTILETTER_ITALIC,EQUATION_NUMBER_CENTER,EQUATION_NUMBER_TABS,CAPTION_ALIGN.TABLE_BORDERSnow resolves the referenced table style'sbasedOnchain (Table Grid draws its borders from styles.xml).EQUATION_UNNUMBERED(FAIL) andEQUATION_NOT_REFERENCED(WARN, 由式 (3-1) 可得).scripts/finalize_docx.py: one mandatory command = mechanical fixes + full audit +DELIVERY GATE: PASS/FAILverdict (non-zero exit on FAIL). SKILL.md and agents/openai.yaml both require a PASS before delivery.Verified against a reproduction of the Codex failure modes (all three now FAIL the gate). 136/136 tests pass; samples regenerated; compliant sample still PASSes.
🤖 Generated with Claude Code