What problem does it solve? Reviewing a branch or PR requires checking both code conventions and spec fidelity, and doing both in one pass causes one axis to mask the other. This Skill separates the two concerns into parallel review agents and produces a persistent, structured report. ## Core Features & Use Cases - Two-axis review: A Standards axis checks the diff against documented conventions (PROJECT.md, CONTEXT.md, qa/REVIEW_RULES.md, ADRs), while a Spec axis verifies the code faithfully implements the originating issue, PRD, or plan. - Parallel sub-agents: Both axes run as isolated sub-agents returning emoji-tagged findings tables (🔴 hard, 🟡 judgement, 💡 nit, ✅ pass), then results are aggregated without cross-axis reranking. - Persistent reports: Writes a standalone markdown report to projects/<name>/qa/reviews/ named by date, branch, and short SHA, with idempotent overwrite at the same commit. - Use Case: After finishing a feature branch, run the review against main to get a verdict table, the single worst blocking issue, and a ship / ship-with-follow-up-ticket / fix-first recommendation. ## Quick Start Ask the AI to review the current branch against main using the project's documented standards and the originating spec.
