Skip to content

title: "Judge — "Is the quality sufficient?"" source: "tasks/TFW-43__research_stage_protocol/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? All 8 AC items verified in verify.md V1-V10. AC-1: exact Mindset wording in all 4 templates. AC-2: no copy in Step 3. AC-3: explicit copy in Step 4. AC-4: FOR EACH in Step 5. AC-5: 🛑 STOP in loop. AC-6: Resume Protocol compatible. AC-7: adapters byte-identical (134 lines, 6097 bytes × 3). AC-8: version 0.8.7 in VERSION + config + CHANGELOG
2 Philosophy aligned P1 (copy-on-enter) → AC-2/3/4. P2 (mindset-first) → AC-1. P3 (STOP between stages) → AC-5. P4 (template = instruction + container) → AC-1 three-layer structure. All 4 HL §7 principles structurally enforced via TS §3 Principles Check table
3 Tech debt documented RF §5 Observations: 1 item — conventions.md L152 doesn't mention Mindset blocks. Genuine low-severity observation
4 Style & standards Naming follows conventions §4. Mindset block format matches review template pattern (map.md L2-3). CHANGELOG follows Keep a Changelog. All files English-only per D29
5 Observations collected 1 observation — style/documentation, not filler. Real issue: conventions.md §4 references stage templates but doesn't describe the Mindset block as a characteristic
6 RF completeness (§6-8) §6 Fact Candidates present — "No fact candidates" with justification (clean AG, no human interaction). §7 Strategic Insights present — "No strategic insights" with justification. §8 Diagrams present — before/after ASCII diagram showing batch-copy vs copy-on-enter flow. All three sections present and substantive

Mode-Specific Checklist

# Check Status Evidence
7 Content quality Mindset blocks are clear, distinct (4 non-overlapping role-nouns on convergent↔divergent/build↔break axes). Test questions are existential-external, stage-specific, escalating (purpose→coverage→insight→robustness). Workflow restructure is clean: dimensional analysis paragraph preserved, FOR EACH is explicit
8 Source verification Key claims traceable: RES D1 (role-nouns) → verified functional mapping. RES D2 (Test format) → existential external confirmed. D31 → file existence semantic restored. D41 → review pattern transferred. All 8 HL knowledge citations resolve to real items (verify.md §Knowledge Citations Verified)

Contradictions with KNOWLEDGE.md

# Knowledge item RF claim Contradiction?
1 D31 (filesystem state machine) Restored by copy-on-enter No — D31 was broken by D50 batch copy, now fixed
2 D50 (research cycle restructure) Step 3 no longer copies templates No contradiction — D50 described folder structure, not copy protocol
3 D41 (4-stage review flow) Pattern transferred to research No — extension, not contradiction

No contradictions found. Changes are additive and consistent with existing architecture decisions.

Checkpoint

Self-check: - [x] Every checklist item has evidence (not just ✅/❌)? - [x] Referenced verify.md findings in DoD assessment? - [x] Checked RF §6-8 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 says "No fact candidates" — justified: AG execution with no human interaction. No challenge needed)

Stage complete: YES