RawinLab Claude Plugins

A collection of Claude Code plugins by RawinLab

---

Quick Install

# Add the marketplace
/plugin marketplace add rawinlab/claude-plugins

# Install any plugin
/plugin install rw-telegram
/plugin install rw-speckit-orchestrator
/plugin install rw-kit

---

Available Plugins

Plugin	Version	Description
rw-telegram	v1.1	Bidirectional Telegram with verbose/summary modes, questions, remote control
rw-speckit-orchestrator	v3.1	Task Tool based Spec-Kit orchestrator with full testing pipeline
rw-kit	v3.0	Fully agentic multi-agent orchestration: dev → UAT → QA pipeline with retry limits and quality gates

---

Plugins

rw-telegram

Bidirectional Telegram integration for Claude Code:

• Smart Notifications - Get notified when Claude completes tasks, needs input, or encounters errors
• Verbose/Summary Modes - Toggle between all events or important only
• Remote Control - Send prompts to Claude directly from Telegram
• Question Forwarding - Answer Claude's questions from your phone
• Project Context - Automatically reads CLAUDE.md before executing tasks
• Auto-start Worker - No manual setup required after configuration

# Setup
/telegram-setup

# Toggle notification mode
/telegram-verbose on    # All events (formatted nicely)
/telegram-verbose off   # Important only (default)

View Documentation

---

rw-speckit-orchestrator

Automated workflow orchestrator for Spec-Kit driven development using Task Tool:

• Task Tool Based - Spawns worker agents for each feature
• Auto-Answer - Workers auto-answer ALL prompts (YES/recommended)
• Context Management - /context + /compact at every level
• Retry Support - Failed workers can be resumed (max 3 times)
• One Feature at a Time - Complete before moving to next

Architecture:

Orchestrator → Task Tool → Worker Agent → Subagents
                              │
                  specify → clarify → plan → analyze → implement

Commands:

/orchestrate --start-from "009"
/orch-status
/orch-stop

View Documentation

---

rw-kit

Fully agentic multi-agent orchestration framework for complete software development lifecycle:

• 10-Phase Pipeline - Clarify → Plan → TodoList → Analyze → Pre-flight → Execute → Test → Quality → UAT → QA Review
• Fully Autonomous - Execute drives entire pipeline from dev through UAT and QA to production-grade
• 5 Custom Agents - SA-Analyst, Tech Lead, Team Lead, Lead Tester, QA Lead
• 12 Commands - Full development lifecycle coverage
• Retry Limits & Escalation - 3 retries per task, BLOCKED/DEGRADED states (no infinite loops)
• Enforcing Quality Gates - Unit, Integration, E2E, Smoke, UAT, QA Review tracked in state
• Auto-Fix Protocols - Pattern-matched fixes for common errors (DI, imports, types)
• Inline UAT + QA - Traceability checks, anti-mock detection, code quality + security review
• 3-Layer Verification - Agent output + File existence + State sync

Pipeline:

Pre-flight → Execute → Tests (enforcing) → Smoke + Auto-Fix → UAT → QA Review → APPROVED

Quick Start:

# Full automated workflow (dev → UAT → QA → production-grade)
/rw-kit:implement user-authentication feature

# Or step-by-step
/rw-kit:clarify requirements/01-auth.md
/rw-kit:plan-module requirements/01-auth-clarified.md
/rw-kit:execute plans/01-1-auth-todolist.md

View Documentation

---

Configuration

rw-telegram

mkdir -p ~/.claude-telegram

cat > ~/.claude-telegram/config.json << 'EOF'
{
  "bot_token": "YOUR_BOT_TOKEN",
  "chat_id": "YOUR_CHAT_ID",
  "allowed_user_ids": [YOUR_USER_ID],
  "worker_port": 37778
}
EOF

chmod 600 ~/.claude-telegram/config.json

rw-speckit-orchestrator

Requires Spec-Kit CLI:

# Install Spec-Kit CLI
uv tool install specify-cli --from git+https://github.com/github/spec-kit.git

# Verify
specify --version

Other requirements:

• jq (for state file parsing)
• gh CLI (for PR creation)

rw-kit

No external dependencies required. State is managed in:

.claude/rw-kit.state.json

Default tech stack (configurable):

• Frontend: Next.js, Shadcn/ui, Zustand
• Backend: NestJS, REST
• Database: PostgreSQL/MySQL, Prisma
• Unit Testing: Jest (80%+ coverage)
• Integration Testing: Jest + Real DB + Seed Data
• E2E Testing: Playwright (user story-driven, seed data)

---

Contributing

Contributions are welcome! Please submit a Pull Request.

---

License

MIT License - see LICENSE for details.

---

Author

RawinLab - github.com/rawinlab