review

Plugin review from marketplace garrettmanley in GarrettManley/claude-marketplace. Commands read from the repo's manifest — 2 days ago.

Skills (1)

• reviewer-personas

  Use when reviewing design documents, PRs, operational docs, work items, wiki pages, or skill files with sub-agent reviewer personas. Provides a reusable library of archetype personas with dispatch protocol and incremental self-improvement.

Slash commands (1)

• /review:review-evolve

  Automate the reviewer-personas Post-Cycle Update Protocol — turn a cycle's Caught/Missed/Hallucinated catches into refined archetype persona files.

Subagents (17)

• accessibility

  Use when reviewing UI design documents, ADO Features, or user stories that describe interactive components, screen layouts, color choices, or dynamic content updates. Catches missing ARIA annotations, keyboard navigation gaps, color contrast violations, and focus management issues.

• api-contract

  Use when reviewing designs or features that introduce or modify API endpoints, route definitions, request/response schemas, or client-facing behavior. Catches idempotency gaps, missing versioning strategies, breaking changes, and inconsistent error schemas.

• architect

  Use when reviewing design documents, ADO Features, or wiki pages for progressive disclosure, mental-model completeness, and claim integrity. Catches undefined concepts, untested quantitative claims, and design-scope leaks.

• architecture-conformance

  Use when reviewing a code diff or PR for dependency direction, earned abstraction, test placement, and duplication of shared-library utilities. Catches imports that point outward, interfaces with no IO boundary and no second implementation, inline test modules, and re-implemented utilities.

• compliance

  Use when reviewing designs that describe regulated data flows, audit trails, certificate lifecycles, or cross-tenant data access. Catches missing evidence of controls, data residency gaps, and undefined audit trail fields.

• contract-surface

  Use when reviewing design documents or features that define configuration defaults, identity/access patterns, external contract surfaces, or SLA claims. Catches prod-as-default paths, imprecise field constraints, and missing compensating controls.

• data-architect

  Use when reviewing design documents, ADO Features, or wiki pages that introduce or modify data models, schema ownership, ingestion pipelines, or multi-tenant data isolation. Focuses on schema design, referential integrity, and data lifecycle.

• ecosystem-context-reviewer

  Use when reviewing designs, PRs, or features that touch external integrations, decision log entries, or partner-facing interfaces. Catches missing integration acknowledgments and ADR conflicts.

• error-handling

  Use when reviewing design documents or features that describe multi-step flows, cross-service interactions, or components with external dependencies. Catches undocumented failure modes, unbounded retries, silent errors at architecture boundaries, and undefined degraded modes.

• incident-commander

  Use when reviewing incident postmortems, operational runbooks, or features with high blast-radius failure modes. Focuses on incident severity, communication clarity, and recovery procedure completeness.

• migration-safety

  Use when reviewing migration procedures, rotation runbooks, slot swap designs, or any multi-step operation that modifies live production state. Catches missing rollback paths, non-idempotent steps, and undefined mixed-version windows.

• new-engineer

  Use when reviewing design documents, wiki pages, or PRs for onboarding clarity and knowledge transfer quality. Catches assumed context, undefined terms, and missing prerequisite steps.

• observability-champion

  Use when reviewing designs that introduce new components, integrations, or data flows without structured logging, metrics, or alerting. Catches observability gaps at design time.

• performance-scalability

  Use when reviewing design documents, ADO Features, or API designs that make latency, throughput, or cost claims. Catches adjective-only performance assertions, missing p99 data, unbounded growth paths, and cost-at-scale gaps.

• security-auditor

  Use when reviewing design documents, PRs, or operational docs that define trust boundaries, access controls, or authentication flows. Focuses on attack surface, least-privilege violations, and defense-in-depth gaps.

• skill-craft-reviewer

  Use when reviewing skill files for Claude Code plugins. Focuses on triggering accuracy, progressive disclosure, Iron Law compliance, and description quality.

• test-strategy

  Use when reviewing design documents, features, or PRs that define or imply a test approach. Catches wrong test level for risk, mock contract gaps, missing boundary conditions, and happy-path-only coverage.

Hooks (1)

• SessionStart

  bash "${CLAUDE_PLUGIN_ROOT}/hooks/session-start-review-nag.sh"

Similar plugins