What problem does it solve? Reviewing large code changes with a single pass often misses issues because one reviewer cannot simultaneously focus on architecture, correctness, and performance. This Skill splits the review into focused parallel sub-reviews and merges the findings into one deduplicated, ranked report. ## Core Features & Use Cases - Scope-Based Mode Selection: Automatically chooses a minimal single-agent review for trivial changes or a full three-agent review for substantive ones. - Parallel Lens Reviews: Launches three concurrent code-reviewer agents covering Design & Architecture, Correctness & Verification, and Quality & Performance. - Aggregated Verdict Report: Deduplicates overlapping findings, ranks them Critical/Important/Minor, preserves file:line references, and produces a merge-readiness verdict. - Use Case: After implementing a feature branch, run this Skill to get a structured multi-perspective review with concrete fixes, then dispatch a backend-developer agent to resolve Critical and Important issues. ## Quick Start Review the changes between BASE_SHA and HEAD_SHA on this branch using parallel focused code reviews and give me an aggregated verdict.
