code-review

<code_review_prompt version="1.0"> <user_input> $1 </user_input>

<inputs_expected>

• Change summary (what/why/risk): {CHANGE_SUMMARY}

• Unified diff or patches: {DIFF}

• Key files (new/modified/deleted): {FILES}

• Build / test logs & coverage: {CI_LOGS} {COVERAGE_SUMMARY}

• Runtime logs or traces (if any): {RUNTIME_LOGS}

• Environment & deployment context: {ENVIRONMENT}

• API schemas/contracts (OpenAPI/GraphQL/Protobuf): {API_SCHEMAS}

• DB schema/migrations: {DB_MIGRATIONS}

• Dependency manifests/lockfiles: {DEPENDENCIES}

• Non-code assets (Dockerfiles, IaC, configs, feature flags): {NONCODE_ASSETS}

• Design notes/ADRs/performance budgets/threat model: {DESIGN_DOCS} </inputs_expected>

• Identify correctness defects, code smells, and anti-patterns.

• Surface exploitable security issues and data-protection risks.

• Spot performance/regression risks and complexity hotspots.

• Check integration points (APIs, events, DB, configs, CI/CD, infra) for compatibility and rollout safety.

• Assess tests for sufficiency, signal, reliability, and coverage.

• Recommend minimal, safe, high-leverage improvements and monitoring.

  <severity_rubric>

• BLOCKER: Exploitable security flaw, data loss risk, broken build/deploy, user-impacting crash, irreversible migration risk, leaked secrets.

• HIGH: Likely prod incident or major regression; authz/auth gaps; significant perf degradation; schema incompatibility.

• MEDIUM: Correctness edge cases; non-exploitable but risky pattern; moderate perf concerns; flaky tests.

• LOW: Maintainability, readability, style, minor test gaps; suggestions.

• NIT: Optional polish. </severity_rubric>

• Scope & Impact: Map all affected files/modules and why each is implicated. Note transitive and runtime impact (build, deploy, config, data).

• Root Cause / Risk Analysis: Explain the change intent, risks introduced, and any hidden assumptions or environmental dependencies.

• Security Review: Use the checklist below; escalate any secret exposure, injection, auth/authz flaws, SSRF/XXE/path traversal, insecure deserialization, command execution, mass assignment, CSRF/XSS, prototype pollution, weak crypto, missing TLS verification, permissive CORS, logging of secrets/PII, dependency vulns, or IaC/container misconfigurations.

• Performance Review: Identify complexity issues, N+1 queries, unbounded loops, memory churn/leaks, blocking I/O on hot paths, missing indexes, cache misuse, chatty network calls, unnecessary allocations/boxing, and concurrency contention.

• Integration Review: Validate API schema changes, versioning, backward/forward compatibility, idempotency, retries/timeouts/circuit breakers, feature-flag rollout, DB migrations (order/locking/rollback), message/event contracts, and config drift.

• Testing Review: Evaluate unit/integration/e2e tests, coverage, negative/property-based cases, concurrency/time-dependent tests, fixture health, determinism, and flakiness risk. Propose a targeted test plan.

• Observability & Operations: Check log levels, PII in logs, correlation/trace IDs, metrics and alerts, runbooks. Recommend what to monitor post-merge.

• Documentation & DX: Flag missing or outdated README/CHANGELOG/ADRs/API docs/config comments/schema diagrams. Note onboarding and maintenance friction.

• Minimal, Safe Fix: Propose the smallest viable change to eliminate blockers/high risks. Include tests and rollout/rollback steps.

  <detailed_checklist>

• Duplicate code / long methods / large classes / deep nesting.

• Leaky abstractions, tight coupling, poor cohesion, improper layering.

• Dead code, unused variables/imports, TODOs that should be addressed now.

• Non-idempotent operations where idempotency is required.

• Edge cases: null/empty/NaN/overflow/encoding/timezone/locale.

• Concurrency: shared state, race conditions, improper locking, async misuse.