By bee-coded
Spec-driven development workflow with TDD enforcement, parallel agent execution, interactive flow control, review gates, and comprehensive code auditing
Add a new phase to the current spec's phases.md and STATE.md
Archive the current spec to .bee/archive/, reset STATE.md, and bump plugin version
Convert audit findings into actionable specs -- generates bee:new or bee:quick tasks from AUDIT-REPORT.md
Run a comprehensive code audit with specialized agents -- security, errors, database, architecture, API, frontend, performance, testing, and end-to-end bug detection
Show diff summary and create a commit with user approval
Audits API design -- endpoint validation, response consistency, error formats, rate limiting, CORS
Audits codebase architecture -- separation of concerns, code organization, duplication, dependency patterns
Traces end-to-end feature flows to find bugs that category-specific auditors miss -- follows data from UI to DB and back
Independently validates audit findings -- reads actual code, verifies evidence, classifies as CONFIRMED / FALSE POSITIVE / NEEDS CONTEXT
Merges validated findings from all audit agents into a structured AUDIT-REPORT.md with severity grouping and recommendations
Core audit knowledge -- severity definitions, finding format, output templates, validation rules. Used by all audit agents.
Context7 MCP usage patterns for framework documentation lookups
BeeDev workflow rules -- TDD mandatory, disk-is-truth, no auto-commit. Spec-driven development with phase lifecycle.
Use when project uses NestJS with RabbitMQ microservices. Covers transport config, message/event patterns, manual ACK, CQRS, hybrid apps, dead letters, health checks, and testing. Detection: check package.json for @nestjs/microservices + amqplib or amqp-connection-manager.
shadcn/ui component library conventions -- use when project has components.json or @/components/ui/ directory. Covers component usage, customization, theming, composition patterns, and common pitfalls.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Uses power tools
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.
Uses Bash, Write, or Edit tools
Uses Bash, Write, or Edit tools
Spec-driven development workflow plugin for Claude Code.
Bee takes a feature idea through a complete, repeatable pipeline: spec > plan > execute > review > test > commit. Every step is orchestrated by Claude Code with review gates between phases. Includes a comprehensive code audit system for analyzing inherited/vibecoded projects.
Option 1: From GitHub (recommended)
# Add the repo as a marketplace source
claude plugin marketplace add https://github.com/BEE-CODED/bee-dev
# Install the plugin
claude plugin install bee
Option 2: Local path (for development)
# Use --plugin-dir flag pointing to the plugin inside the repo
claude --plugin-dir /path/to/bee-dev/plugins/bee
/bee:init # Detect stack, create .bee/ config
/bee:new-spec # Conversational discovery → spec → phases
/bee:plan-phase 1 # Plan phase 1 with tasks and wave grouping
/bee:execute-phase 1 # Execute with parallel TDD implementation
/bee:review # 4-step code review pipeline
/bee:test # Generate test scenarios, verify with developer
/bee:commit # Guided commit with diff summary
/bee:quick # Fast-track: describe, execute, commit (no spec needed)
/bee:audit # Comprehensive 9-agent code audit (great for inherited projects)
| Command | Description |
|---|---|
/bee:init | Initialize project -- detect stack, create .bee/ config |
/bee:progress | Show current project state and suggest next action |
/bee:resume | Restore full context after a break |
/bee:compact | Smart compact -- preserve bee context, then compress conversation |
/bee:memory | View accumulated agent memories for the current project |
/bee:refresh-context | Re-run codebase context extraction, overwriting CONTEXT.md |
/bee:create-agent | Create a custom project-local agent extension |
/bee:create-skill | Create a custom project-local skill extension |
/bee:update | Update bee statusline and clean up legacy local copies |
| Command | Description |
|---|---|
/bee:new-spec | Create feature spec through conversational discovery. --amend to modify, --from-discussion to use discussion notes |
/bee:plan-phase N | Plan a phase: task decomposition, research, wave assignment + auto-fix review loop |
/bee:plan-review N | Review a phase plan (TASKS.md) against the spec before execution |
/bee:add-phase | Append a new phase to the current spec |
/bee:discuss | Guided brainstorming-style codebase-grounded discussion before creating a spec |
| Command | Description |
|---|---|
/bee:execute-phase N | Execute a planned phase with wave-based parallel TDD agents |
/bee:quick | Fast-track task. --fast for direct mode, --amend N to modify existing, --review for post-review |
| Command | Description |
|---|---|
/bee:review | Multi-agent parallel review (4 specialists) with finding validation, escalation, and auto-fix. --loop for auto-loop |
/bee:review-implementation | Context-aware review -- full spec mode (4 agents per stack) or ad-hoc mode (3 agents) |
/bee:fix-implementation | Standalone fix -- reads review output and fixes confirmed findings sequentially |
/bee:test | Generate manual test scenarios and verify with developer |
/bee:test-e2e | Generate and run Playwright E2E tests with Page Object Model |
| Command | Description |
|---|---|
/bee:audit | Comprehensive 9-agent code audit with finding validation and structured report. --only security,database for selective |
/bee:audit-to-spec | Convert audit findings into actionable specs by severity. --critical for critical only, --dry-run to preview |
| Command | Description |
|---|---|
/bee:commit | Show diff summary, suggest commit message, require confirmation |
/bee:archive-spec | Archive completed spec, reset STATE.md, bump plugin version |
/bee:eod | End-of-day integrity check with 4 parallel audits |
/bee:init → /bee:new-spec → /bee:plan-phase 1 → /bee:execute-phase 1 → /bee:review → /bee:test → /bee:commit
↑ repeat for each phase ↓
/bee:review-implementation → /bee:archive-spec → /bee:eod
/bee:quick fix the login button alignment # Agents mode (default): research + implement
/bee:quick --fast fix the button color # Direct execution, no agents
/bee:quick --amend 3 # Modify existing quick task plan #3
/bee:quick --review update the footer links # Execute + review before commit
npx claudepluginhub bee-coded/bee-dev --plugin beePersona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code
Skills-first specification-driven development framework with 7 agent skills for planning, implementation, review, and shipping. Natural language activation with intelligent agent orchestration. Includes /plan, /implement, /research commands plus managing-specifications, implementing-features, and reviewing-and-shipping skills.
Spec-driven development pipeline for Claude Code: interview-to-spec, spec-approach, tdd, verify, diagnose, and a feature orchestrator.
Specification-Driven Development with Process Discipline for Claude Code
Spectacular skills/commands library for spec-anchored development and parallel task execution
Specification-driven development workflow: specify → plan → tasks → implement