Generate and run check artifacts for a spec. Self-sufficient: MCP preflight, 2-agent generation, автоматически по всем Requirements, auto state transition.
Реализовать задачи из tasks.md текущего change'а с встроенной L1/L2/L3 верификацией. Читает test-plan.md как контекст. Обновляет .sdd-state.yaml на каждом подшаге. Обновляет openspec/specs/index.yaml для новых capabilities (только при verify-ok).
Update dev: namespace to latest version from awesome-claude. Checks dependencies and updates them if needed.
Create a well-structured git commit following project conventions. Analyzes staged and unstaged changes, drafts a detailed commit message with What/Why/Details sections, and commits. Auto-detects task ID format from git history.
Bug fixer: traces root cause via /tracing (analysis only), then fixes via /tdd (test-first implementation). Combines incident analysis with TDD-driven repair. Use when: something is broken, a feature doesn't work, or a bug needs fixing.
Initialize a DDD monorepo with backend (FastAPI, SQLAlchemy, multi-BC architecture) and frontend (React 19, Vite, TypeScript) packages. Scaffolds full project structure: root orchestration (Makefile, docker-compose), backend with one bounded context (domain entity, repository, use case, infra repo, routes, schemas), frontend with API client, router, UI kit seed, and test infrastructure for both (pytest 6-type strategy, vitest, architecture tests). Ready to `make check` after generation.
TDD workflow: writes tests FIRST, runs them (red), writes implementation (green), refactors. Works as a senior QA automation engineer with 20 years of experience. Covers all test layers: unit, state, security, cases, integration, e2e, contract. Always follows critical path first, then corner cases. Use when implementing features, fixing bugs, or adding test coverage via TDD.
Tracing & incident analyzer: traces why a feature doesn't work after deploy, locates the root cause in code across all layers (frontend → API → backend → DB), generates PlantUML sequence + C4 component diagrams showing the failure path, produces a structured issue table, and assesses implementation risks. Use when: a deployed feature doesn't work, triaging bugs, incidents, or production issues.