review-synthesis

Consolidate multi-pass code review findings into a verified single report.

Updated Mar 16, 2026
One-click install
npx skills add https://github.com/liminal-ai/code-steward --skill review-synthesis-liminal-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-synthesis
Source: https://github.com/liminal-ai/code-steward/tree/main/.claude/plugins/code-steward-reviews/skills/review-synthesis
Command: npx skills add https://github.com/liminal-ai/code-steward --skill review-synthesis-liminal-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill aggregates outputs from multiple review passes (standard, security, performance, architecture, test-coverage, silent-failure), verifies each finding against the actual code, deduplicates conflicting results, and produces a final consolidated report that tech leads can act on.

Core Features & Use Cases

  • Verification across reviews: Cross-checks findings from standard, security, performance, architecture, test-coverage, and silent-failure reviews against the actual code.
  • Deduplication and conflict resolution: Merges duplicates and resolves contradictory findings to present a coherent final report.
  • Contextual fidelity: Preserves per-review data fields and ties findings to repository context (files, lines, metadata).

Quick Start

Synthesize review outputs and verify every finding against the code to produce a consolidated, unimpeachable final report.

Frequently Asked Questions about review-synthesis

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I consolidate multiple code review outputs into a single report?

You consolidate multiple code review outputs by cross-checking findings from standard, security, performance, architecture, test-coverage, and silent-failure passes against the actual codebase to produce one verified report.

What is code review synthesis and when do I need it?

Code review synthesis aggregates multi-pass review outputs, verifies each finding against the repository code, and deduplicates conflicting results to produce a coherent final report for tech leads.

Can I deduplicate conflicting code review findings from different passes?

Yes, you can deduplicate conflicting findings by merging duplicate issues and resolving contradictory results across standard, security, performance, and architecture reviews to present a coherent final report.

Does code review verification preserve per-review metadata and repository context?

Yes, code review verification preserves per-findings fields from each review pass and ties verified issues to repository context including specific files, lines, and metadata.

What is the best way to verify code review findings against the actual codebase?

The best way to verify code review findings is reading repository files directly to cross-check each finding against the actual code, ensuring only verified issues are reported.

Why should I verify code review findings before generating a final report?

You should verify code review findings to eliminate false positives and ensure only verified issues from standard, security, performance, architecture, test-coverage, and silent-failure passes are reported.