Skip to content

title: "Judge — "Is the quality sufficient?"" source: "tasks/TFW-47__codex_adapter_shortcut_skills/phase-a/review/judge.md"


Judge — "Is the quality sufficient?"

Mindset: Judge. You have the evidence from Verify. Now rule on quality. Every ✅ needs proof. Every ❌ needs a specific finding. Test: "Would I stake my reputation on this passing production review?" Mode: docs Verify findings: verify.md

Universal Checklist

# Check Status Evidence
1 DoD met? verify.md: All 6 ACs verified. V1-V7 all ✅ match. No discrepancies found.
2 Philosophy aligned HL §7 P1 (mandatory, not optional): conventions.md §3/§4 use "MUST contain" language. P2 (template reflects values): EV template has structured Environment header + per-AC table + Verdict — trace-first, reproducible.
3 Tech debt documented RF §6 has 3 Observations: compilable contract Source Manifest, project_config.yaml templates, review template RF §5 references. All are real issues — naming/config gaps.
4 Style & standards EV naming follows D28 (2-letter abbreviation, consistent with HL/TS/RF). File naming in conventions.md §4 table follows existing patterns. New sections use same heading hierarchy as existing content.
5 Observations collected 3 observations, all typed "naming." Quality: Observation #1 (compilable_contract.md Source Manifest) is a real gap. #2 (project_config.yaml) correctly notes consistency with research templates. #3 (review/verify.md) correctly assessed as no-action. All are substantive, not filler.
6 RF completeness (§7-9) §7 Fact Candidates: "No fact candidates." — section present, empty content valid. §8 Strategic Insights: "No strategic insights." — section present, empty content valid. §9 Diagrams: BEFORE/AFTER ascii diagram present showing evidence pipeline change.
7 Evidence completeness All 6 TS AC items had Evidence: N/A — definition documents, no runtime verification applicable. EV file exists with correct all-N/A verdict. This is the correct treatment for methodology changes (templates, conventions, workflows).

Mode-Specific Checklist (docs)

# Check Status Evidence
7 Content quality EV template is clear, well-structured, and self-documenting. Conventions updates use consistent language ("MUST contain"). TS/RF template changes are minimal and precise. Handoff substeps are actionable (create → copy → fill → walk → verdict → attachments).
8 Source verification D53 correctly cites "0/38 tasks created evidence/" from HL §2. EV template structure aligns with iter1 RES decisions D1-D7. Anti-pattern language at L442 correctly references evidence/ folder.

Contradictions with KNOWLEDGE.md

# Knowledge item RF claim Contradiction?
1 D52 (Evidence Layer — TFW-46) D53 extends, doesn't modify D52 No — D52 preserved as original concept, D53 = enforcement extension
2 D16 (evidence/ optional) D53 revokes D16 No — explicit revocation with justification, not silent overwrite

Checkpoint

Self-check: - [x] Every checklist item has evidence (not just ✅/❌)? - [x] Referenced verify.md findings in DoD assessment? - [x] Checked RF §7-9 for presence AND quality (not just existence)? - [x] KNOWLEDGE.md cross-referenced — contradictions documented or "None"? - [x] Fact Candidates from RF reviewed — any that need challenge? - RF §7 says "No fact candidates." — acceptable for a methodology phase that follows well-researched decisions (iter1 D1-D7).

Stage complete: YES