Enhance review template transparency and generation quality by adding detailed scoring rubrics, deduction traceability, and a structured challenge response protocol.

Summary

This PR addresses core user feedback about opaque and inconsistent review scoring by adding transparent scoring rubrics to all 5 review templates, and quality targets to 2 generation
templates. Users can now trace every score to specific rubric criteria and deduction details, and the system handles score challenges objectively through a defined 3-step protocol.

Changes

Key Modifications

Technical Details

Review Templates (5 files) — Each template now includes:

1. Scoring Rubrics: Per-category rubric tables with 4 score ranges (90-100, 70-89, 50-69, Below 50) and typical deductions with specific point values
2. Enhanced Scoring Breakdown: Added "Rubric Basis" and "Deduction Details" columns to the scoring table for full traceability
3. Suggestion Score Cap Rule: Suggestions/Nice-to-Have items limited to max 5 points impact; resolving all Critical + Warning issues guarantees ≥ 95
4. Score Validation Checklist: 6-point verification (deduction-issue matching, arithmetic, cap compliance, consistency)
5. Score Challenge Response Protocol: 3-step process — Provide Evidence → Ask for Specifics → Targeted Re-evaluation

Generation Templates (2 files) — Each template now includes:

• Quality Targets: Per-category checklists aligned with the corresponding review template's rubric categories, with a self-check reminder before saving

Testing

Test Coverage:

• All 5 review templates contain "Scoring Rubrics" section
• All 5 review templates contain "Rubric Basis" and "Deduction Details" in Scoring Breakdown
• All 5 review templates contain "Suggestion Score Cap Rule"
• All 5 review templates contain "Score Validation Checklist"
• All 5 review templates contain "Score Challenge Response Protocol"
• Both generation templates contain "Quality Targets" section
• Quality Target categories match corresponding review rubric categories
• Consistent format across all templates (Score Range/Criteria tables, Typical Deductions lists)
• Constitution-absent edge case handled (defaults to 100 with weight redistribution)

How to Verify

1. Checkout the branch: git checkout 2026-0409-1149ob-enhance-review-quality
2. Review any review template (e.g., templates/commands/review-spec.md) and verify:
   • "Scoring Rubrics" section with per-category tables
   • "Scoring Breakdown" table has "Rubric Basis" and "Deduction Details" columns
   • "Suggestion Score Cap Rule" section
   • "Score Validation Checklist" section
   • "Score Challenge Response Protocol" section
3. Review generation templates (spec-to-plan.md, plan-to-tasks.md) for "Quality Targets" section
4. Install and test: uv tool install --force .
5. Run a review command (e.g., /codexspec:review-spec) on a sample spec to verify the new sections appear in the report

Checklist

• Code follows project style guidelines
• Self-review completed
• Only templates/commands/ modified (per constitution)
• All pre-commit hooks pass (markdownlint, etc.)
• Consistent format across all 5 review templates
• Generation quality targets aligned with review rubrics

Notes

• No breaking changes — existing review reports will now include additional sections
• Template length increased (~120-140 lines per review template) to accommodate rubrics and protocols
• The language.output configuration continues to work — all new sections are in English templates but will be dynamically translated by Claude at runtime