← ClaudeAtlas

source-code-reviewlisted

Runs a holistic senior-engineer source-code review of production and test code against spec/project/source-code-review/ plus its frontend extension spec/frontend/source-code-review/, and persists a severity-classified report whose disjoint work packages are parallel-dispatchable to specialists. The default `review` operation detects each language and surface and dispatches the matching read-only reviewer agent (`python-code-reviewer`, `frontend-code-reviewer`; anything unprofiled is reported unsupported), then persists the report under .audits/source-code-review/. The `plan` operation hands that report to implementation-plan-author. Invoke to review a codebase like an experienced developer, find domain-knowledge duplication, or find business logic leaking into the client; also German. Don't use for the deep OWASP audit (code-security-reviewer), UX judgment (frontend-usability-optimizer), single-tier test conformance (tier reviewers), or to run lint and tests (quality-gate). Supports resume.
nolte/claude-shared · ★ 0 · Code & Development · score 74
Install: claude install-skill nolte/claude-shared
# Source Code Review Run a holistic senior-engineer source-code review over production **and** test code, persist one severity-classified report, and hand its work packages to specialists for parallel remediation. The default `review` operation detects and dispatches; the `plan` operation turns the persisted report into a specialist-mapped implementation plan. **There is no apply** — remediation is real code work owned by specialists, never an in-place rewrite from this skill. Implements `spec/project/source-code-review/` — the spec defines the language-agnostic core dimensions (D1–D10), the language profiles, the tooling-first rule, and the report contract with disjoint work packages — plus its **frontend surface extension** `spec/frontend/source-code-review/`, which overlays the frontend dimensions (F1–F11), the component-slice review unit, the framework profiles, the implementation-proposal requirement, and the delimitation from the UX review. This skill binds those rules to the on-disk procedure and owns detection, persistence, and the plan handover; the review judgment itself lives in the dispatched reviewer agent. ## German trigger phrases This skill also triggers on equivalent German-language requests, including: - "Code-Review durchführen" / "Source-Code-Review wie ein erfahrener Entwickler" - "fachliche Duplizierung finden" / "doppelte Geschäftslogik aufspüren" - "Testcode-Qualität reviewen" / "Produktiv- und Testcode prüfen" - "Frontend-Components reviewen" / "