← ClaudeAtlas

reviewlisted

Use after implementation to run independent code review - dispatches reviewer agents based on task level, ensures author-reviewer separation
sdsrss/sgc · ★ 0 · AI & Automation · score 72
Install: claude install-skill sdsrss/sgc
# Review Dispatch reviewer agents (fresh context) against the current task's diff. Reviewers cannot read `solutions/` — this is Invariant §1, not a suggestion. **Core principle:** the author cannot review their own work. Reviewers must judge without historical-solution bias. ## When to Use - User runs `/review` after `/work` is complete - Before `/ship` (a code review is required for L1+) ## Permission | Directory | Access | |-----------|--------| | decisions | R | | progress | R | | solutions | **FORBIDDEN** (§1) | | reviews | W | ## Routing - **Behavior**: [`src/commands/review.ts`](../../../../src/commands/review.ts) (`runReview`) - **Base reviewer**: [`src/dispatcher/agents/reviewer-correctness.ts`](../../../../src/dispatcher/agents/reviewer-correctness.ts) — runs at every level - **L3 specialists**: [`src/dispatcher/agents/reviewer-specialists.ts`](../../../../src/dispatcher/agents/reviewer-specialists.ts) — `reviewer.{security,migration,performance,infra}` spawn in parallel when the diff matches their trigger keywords. Aggregate verdict = worst-of (`pass < concern < fail`) - **Manifest**: 9 reviewers in [`contracts/sgc-capabilities.yaml`](../../../../contracts/sgc-capabilities.yaml) (correctness/security/performance/tests/maintainability/adversarial/spec/migration/infra). `tests`/`maintainability`/`adversarial`/`spec` are manifested as forward-references — not yet wired into `runReview` - **Scope pin**: `spawn.ts` emits `scope_tokens:` + `FORBIDDEN from: read:so