← ClaudeAtlas

review-pro-synthesizelisted

Stage 3 of review-pro: dedup, weight, resolve conflicts, calibrate severity, and produce the final verdict + report from specialist findings. Use to merge reviewer results into one review verdict.
tufantunc/review-pro · ★ 4 · Code & Development · score 74
Install: claude install-skill tufantunc/review-pro
# Review-Pro Synthesis (Stage 3) You are the orchestrator's final stage. You receive the structured findings from all dispatched reviewers and produce ONE unified review. ## Steps 1. **Collect** all finding blocks from the dispatched reviewers. 2. **Dedup** by `(file, line±5, category-root, overlap_hints)`: the same issue flagged by multiple reviewers collapses into one. 3. **Weight:** a finding flagged by ≥2 reviewers gets a conviction boost — annotate it "flagged by N reviewers". 4. **Resolve conflicts** by ownership — the domain owner sets severity (see table). 5. **Calibrate severity:** enforce the anti-overreporting bar. Downgrade anything not fully traced to evidence. Never upgrade beyond what a specialist justified. 6. **Verdict** + prioritized findings + remediations. ## Conflict ownership | Domain | Severity authority | |---|---| | security / auth / secrets | security-reviewer | | data integrity / migrations | db-reviewer | | contract / back-compat | api-contract-reviewer | | maintainability / structure | craft-reviewer | | performance | performance-reviewer | | test correctness | tests-reviewer | | accessibility | a11y-reviewer | ## Verdict (see core/shared/severity.md) - **BLOCK:** any unaddressed Critical or High. - **REQUEST CHANGES:** any Medium or above. - **APPROVE:** only Low/Nitpick, or no findings. ## Output A markdown report. Lead with the verdict and Critical/High. Do not restate raw specialist dumps — present the unified, deduped view. ``` ## Verdi