By kirkchen
Orchestrates a full BDD workflow from exploration to archive using Gherkin feature files, with TDD-based implementation in git worktree isolation, automated task breakdown, multi-agent validation, and living documentation sync.
Use when implementing a Beat change — requires gherkin or proposal artifact to be done first
Use when a Beat change is complete (implemented, or distilled and verified) and ready to archive — not for verifying implementation
Use when starting a Beat change to create spec artifacts — not for task breakdown, implementation, or exploration
Use when extracting BDD specs from existing code — for adopting Beat in an established codebase or distilling a module into feature files
Use when thinking through ideas, investigating problems, or clarifying requirements — before or during a Beat change
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
English | 繁體中文
Think first, code second. Beat is a Claude Code plugin that makes you write Gherkin scenarios before touching code — then drives TDD implementation from those specs.
You: "Add user login"
Claude: *writes 400 lines, misses rate limiting,
tests are an afterthought, edge cases surface in PR review*
You: /beat:design add-user-login
Beat generates:
┌─────────────────────────────────────────────────┐
│ Feature: User Login │
│ │
│ Scenario: Successful login with valid creds │
│ Scenario: Invalid password shows error │
│ Scenario: Account locked after 5 attempts │
│ Scenario: Session expires after 30 min idle │
│ Scenario: Login from new device sends email │
└─────────────────────────────────────────────────┘
You: /beat:apply
Beat implements each scenario with TDD:
✗ Write test for "Account locked after 5 attempts" (red)
✓ Implement lockout logic (green)
✓ Refactor (clean)
→ Next scenario...
Every scenario gets a test. Every test links back to the spec. Nothing slips through.
From the plugin marketplace:
/install kirkchen/beat
Or locally:
claude --plugin-dir /path/to/beat
On Codex:
codex plugin marketplace add https://github.com/kirkchen/beat
Then in the Codex UI plugins panel, install Beat. See docs/INSTALL-CODEX.md for the full flow and the AGENTS.md snippet that replaces the SessionStart hook (Codex doesn't support hooks).
Then, in your project:
/beat:setup # Detects your stack, creates beat/config.yaml
For most changes, this is all you need:
/beat:design fix-expired-session # Describe the behavior change → generates Gherkin
/beat:apply # TDD: test → implement → next scenario
/beat:archive # Sync features to living docs, clean up
For complex changes, add /beat:verify before archive — it dispatches an independent agent to validate your implementation against the specs, catching gaps that the implementer's own context would miss.
Beat scales up when you need it (see Full Pipeline), but the simple path works for everyday fixes and small features.
Most BDD tools only work for new code. Beat works backwards too — extract Gherkin from your existing codebase:
/beat:distill src/billing/ # Reads your code, generates feature files
@distilled @behavior @happy-path
Scenario: Monthly billing adjusts for short months
Given a subscription billing on the 31st
When February billing cycle runs
Then the charge date adjusts to the 28th
Now you have living documentation of what your system actually does. Next time you change billing logic, Beat knows what behavior exists and what tests to write.
This is the recommended entry point for existing projects. Distill first, then use the full pipeline for future changes.
explore → design → plan → apply → verify → archive
| Command | What it does | When to use |
|---|---|---|
/beat:explore | Think through ideas, no code | Unclear requirements, brainstorming |
/beat:design | Create change + generate specs | Starting any change |
/beat:plan | Task breakdown + multi-role review | Complex features (5+ scenarios) |
/beat:apply | TDD implementation per scenario | Every change |
/beat:verify | Independent verification against specs | Before shipping complex changes |
/beat:archive | Sync features + archive change | After every change |
Pick your path by size:
| Change size | Commands | Example |
|---|---|---|
| Bug fix | design → apply → archive | Fix off-by-one in date calc |
| Feature | design → apply → verify → archive | Add password reset flow |
| Large feature | design → plan → apply → verify → archive | Payment processing system |
Every change lives in beat/changes/<name>/ with a status.yaml tracking where you are.
Each change can include these artifacts (you choose which):
| Artifact | Default | Purpose |
|---|---|---|
features/*.feature | Included | Gherkin scenarios — the spec |
proposal.md | Optional | Why this change exists |
design.md | Optional | Technical decisions |
tasks.md | Optional | Implementation plan |
For purely technical changes (refactoring, tooling, deps), you can skip Gherkin entirely and drive from proposal.md instead.
npx claudepluginhub kirkchen/beat --plugin beatAI coding lifecycle skills: triage-issue / investigate-issue (intake) → self-review (pre-push) → pr-review (on-PR, 4-role subagent dispatch) → pr-babysit (until merge).
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review