code-review · git:20261118.4ed259f · 2026-11-18 · sha256 309c4a656611703f
code-review git:20261118.4ed259fA
Immutable. This exact content never changes and is served at /api/v1/blob/309c4a656611703f.
No tool to install first
claude plugin marketplace add https://markdownregistry.com/mp/a/art_gfo4vuwbo7kywoqr/309c4a656611703f2f457abfffbbc1ad092f9a34a7cbeb5e4af7d288f81a1b1e.json
Then /plugin install code-review@markdownregistry-gfo4vuwbo7kywoqr in Claude Code. It installs this exact version, the one graded A here, pinned to the commit it was graded at, and leaves it there.
--- name: code-review description: Audits diffs, branches, or completed tickets across specification fidelity, configurability, structured logging, security/credential hygiene, performance optimization, shared UI/DRY reuse, modularity, and test integrity. Use after implementing a ticket or reviewing a PR. --- # Code Review (Industrial 2-Axis & Post-Ticket Audit) Audit the diff (`git diff` or branch diff) against **Specification Fidelity** and **Industrial Engineering Standards** using [references/review-rubric.md](references/review-rubric.md). ## Decision tree 1. **Is the target diff (`git status` / `git diff`) resolved and non-empty?** - **No** → Inspect `git status` and `git diff` via `run_command`, or ask the user for the base ref. - **Yes** → Proceed to step 2. 2. **Are you reviewing your own ticket implementation or a multi-module diff?** - **Yes** → Run `node .agents/skills/implement/scripts/audit-diff.mjs` first, then delegate read-only audits via `invoke_subagent` to `adversarial-reviewer` (and `security-auditor` / `architecture-planner` when auth, secrets, schemas, or module boundaries change). - **No** → Execute the review workflow below directly. ## Review workflow 1. **Inspect Diff & Sibling Callers**: - Capture `git diff` via `run_command`. - For every modified exported symbol, widget/component prop, or schema, run `rg -n` to confirm unchanged callers remain valid. 2. **Audit Against [references/review-rubric.md](references/review-rubric.md)**: - Specification & ambiguity fidelity (zero invented endpoints/fields/assumptions). - Configurability & structured logging (`traceId`, `<REDACTED>` secrets/PII, zero hardcoded URLs/magic constants, zero stray `print`/`console.log`). - Security & credential leak hygiene (zero embedded secrets, BOLA/IDOR `tenant_id` scoping, atomic locks, SSRF/path guards). - Performance & resource lifecycle (zero N+1, optimal data structures, virtualization/pagination, isolate/worker offload, `dispose()`/`close()`). - Shared UI & DRY reuse (`Reuse → Extend → Compose → Create`, design tokens). - Modularity, maintainability & test integrity (zero `.skip`, `skip: true`, SUT mocks, weak assertions, or production fakes). 3. **Report Ranked Findings**: - Group by `🔴 Blocking`, `🟡 Structural`, and `🟢 Nit` with exact verified `file:line` citations and concrete fixes. - Conclude with **`APPROVE`**, **`APPROVE WITH NITS`**, or **`REQUEST CHANGES`**.