II.
Page JSON
Structured · livepage:docs-reference-getting-started
Getting Started with Babysitter Plugin json
Inspect the normalized record payload exactly as the atlas UI reads it.
{
"id": "page:docs-reference-getting-started",
"_kind": "Page",
"_file": "wiki/docs/reference/getting-started.md",
"_cluster": "wiki",
"attributes": {
"nodeKind": "Page",
"sourcePath": "docs/reference/GETTING_STARTED.md",
"sourceKind": "repo-docs",
"title": "Getting Started with Babysitter Plugin",
"displayName": "Getting Started with Babysitter Plugin",
"slug": "docs/reference/getting-started",
"articlePath": "wiki/docs/reference/GETTING_STARTED.md",
"article": "\n# Getting Started with Babysitter Plugin\n\n> Get your first orchestrated workflow running in under 20 minutes.\n\nThis guide walks you through installing the Babysitter plugin and running your first successful orchestration. By the end, you will understand how to use the Babysitter skill to manage complex, multi-step workflows.\n\n---\n\n## Table of Contents\n\n1. [Prerequisites](#prerequisites)\n2. [Quick Installation (5 minutes)](#quick-installation-5-minutes)\n3. [First Orchestration (10 minutes)](#first-orchestration-10-minutes)\n4. [Understanding the Output](#understanding-the-output)\n5. [Troubleshooting](#troubleshooting)\n6. [Next Steps](#next-steps)\n\n---\n\n## Prerequisites\n\nBefore installing, verify your system has the required dependencies.\n\n### Required Software\n\n| Software | Minimum Version | Check Command |\n|----------|-----------------|---------------|\n| Node.js | v18+ | `node --version` |\n| npm | Any recent | `npm --version` |\n| Git | Any recent | `git --version` |\n| jq | Any version | `jq --version` |\n\n### Verify Prerequisites\n\nRun the verification script to check all dependencies at once:\n\n```bash\n# From your project directory containing the plugin\n```\n\n**Expected output when all checks pass:**\n\n```\nBabysitter Plugin Installation Verification\n============================================\n\n=== Environment Checks ===\n✓ Environment: Linux/macOS/Windows detected\n\n=== Required Dependencies ===\n✓ Node.js: Version 20.x.x (>=18 required)\n✓ npm: Version 10.x.x\n✓ git: Version 2.x.x\n✓ jq: Version 1.x\n\n=== Plugin Checks ===\n✓ Plugin Structure: All required directories and files present\n✓ SDK CLI: Version x.x.x\n\n=== Summary ===\nAll checks passed! The babysitter plugin is ready to use.\n```\n\n### Install Missing Dependencies\n\n**Node.js and npm:**\n- Download from [nodejs.org](https://nodejs.org/) (v18 LTS or later recommended)\n\n**jq (required for hooks):**\n```bash\n# macOS\nbrew install jq\n\n# Ubuntu/Debian\nsudo apt-get install jq\n\n# Windows (using Chocolatey)\nchoco install jq\n```\n\n---\n\n## Quick Installation (5 minutes)\n\nChoose one of three installation methods:\n\n### Option A: Via Claude Code Marketplace (Recommended)\n\nIn Claude Code, run:\n\n```\n/plugin marketplace add a5c-ai/babysitter-claude\n/plugin install babysitter@a5c.ai\n```\n\nAfter installation:\n1. Restart Claude Code if prompted\n2. Verify with `/skills` - look for \"babysitter\" in the list\n\n### Option B: Manual Installation from Source (Global)\n\n```bash\n# Clone the source repo\ngit clone https://github.com/a5c-ai/babysitter.git ~/src/babysitter\ncd ~/src/babysitter\n\n# Generate the harness-specific plugin bundles from plugins/babysitter-unified\nnpm run generate:plugins\n\n# Install the generated Claude Code bundle\nmkdir -p ~/.claude/plugins/babysitter\ncp -r artifacts/generated-plugins/claude-code/* ~/.claude/plugins/babysitter/\n```\n\n### Option C: Manual Installation from Source (Project-Local)\n\nFor project-specific plugin usage:\n\n```bash\n# Generate the plugin bundle from the repo root first\nnpm run generate:plugins\n\n# Create plugin directory in your project\nmkdir -p .claude/plugins\n\n# Copy the generated Claude Code bundle into the project-local plugins dir\nmkdir -p .claude/plugins/babysitter\ncp -r artifacts/generated-plugins/claude-code/* .claude/plugins/babysitter/\n```\n\n### Verify Installation\n\nIn Claude Code, verify the skill is loaded:\n\n```\n/skills\n```\n\nYou should see `babysitter:babysit` and `babysitter:call` in the list.\n\n---\n\n## First Orchestration (10 minutes)\n\nThe Babysitter plugin is used through **skill invocation** - you ask Claude to use the babysitter skill, and Claude handles all the orchestration internally.\n\n### Method 1: Direct Skill Invocation\n\nUse the slash command to invoke the skill:\n\n```\n/babysitter:call implement a Hello World function with TDD targeting 80% quality\n```\n\n**Other modes available:**\n\n| Mode | Command | Behavior |\n|------|---------|----------|\n| Interactive | `/babysitter:call` | Pauses at breakpoints for approval (default) |\n| YOLO | `/babysitter:yolo` | Full auto, no breakpoints — ship while you sleep |\n| Forever | `/babysitter:forever` | Never-ending loop for monitoring/periodic tasks |\n| Plan | `/babysitter:plan` | Creates process without executing — review first |\n\n### Method 2: Natural Language Request\n\nSimply ask Claude to use the babysitter skill:\n\n```\nUse the babysitter skill to implement a Hello World function.\nRequirements:\n- Create a function that returns \"Hello, World!\"\n- Function accepts an optional name parameter\n- Returns \"Hello, {name}!\" when name is provided\n- Include unit tests\n- Use TDD methodology\n```\n\n### What Happens Next\n\nWhen you invoke the skill, Claude will:\n\n1. **Interview Phase** - Claude may ask clarifying questions about your requirements, scope, and preferred methodology\n\n2. **Process Creation** - Claude creates a custom process tailored to your request\n\n3. **Run Setup** - Claude initializes the orchestration run (you'll see a message about the run being activated)\n\n4. **Iteration Loop** - Claude executes the workflow step by step:\n - Executes tasks (code generation, testing, refactoring)\n - Posts results to the run\n - Handles quality gates automatically\n\n5. **Breakpoints** - If human approval is needed, Claude will ask you directly (in interactive mode) before continuing\n\n6. **Completion** - When done, Claude reports the final results\n\n### Example Session\n\nHere's what a typical session looks like:\n\n```\nYou: /babysitter:call create a calculator module with add and subtract functions using TDD\n\nClaude: I'll orchestrate this using the babysitter skill. Let me first understand your requirements...\n\nClaude: What programming language would you prefer?\n- TypeScript (Recommended)\n- JavaScript\n- Python\n\nYou: TypeScript\n\nClaude: 🔄 Babysitter run activated in this session!\nCreating process for TDD calculator implementation...\n\n[Claude creates the process and begins orchestration]\n\nClaude: Phase 1: Writing failing tests for add function...\n[Shows test code]\n\nClaude: Phase 2: Implementing add function to pass tests...\n[Shows implementation]\n\nClaude: Phase 3: Quality gate - Tests passing: ✓\nCurrent quality score: 75%\n\n[Continues iterating until quality threshold is met]\n\nClaude: ✅ Run completed successfully!\n- All tests passing\n- Quality score: 85%\n- Files created: src/calculator.ts, tests/calculator.test.ts\n```\n\n### Choosing a Methodology\n\nYou can specify which methodology to use:\n\n```\n/babysitter:call use GSD methodology to quickly prototype a REST API\n\n/babysitter:call use TDD Quality Convergence to implement user authentication\n\n/babysitter:call use Spec-Kit for implementing the requirements in spec.md\n```\n\nSee the [Process Selection Guide](https://github.com/a5c-ai/babysitter/blob/main/docs/reference/PROCESS_SELECTION.md) for help choosing the right methodology.\n\n---\n\n## Understanding the Output\n\n### Run Directory Structure\n\nBabysitter creates a run directory with all artifacts:\n\n```\n~/.a5c/runs/<runId>/ # default; use <repo>/.a5c/runs/<runId>/ when BABYSITTER_RUNS_SCOPE=repo\n├── run.json # Run metadata\n├── inputs.json # Your inputs\n├── journal.jsonl # Event log (append-only)\n├── state.json # Current state cache\n├── code/ # Process implementation\n│ └── main.js # Process entry point\n├── tasks/ # Task execution artifacts\n│ └── <effectId>/\n│ ├── input.json\n│ ├── result.json\n│ ├── stdout.log\n│ └── stderr.log\n└── artifacts/ # Human-readable outputs\n └── *.md\n```\n\n### Iteration Status Values\n\n| Status | Meaning | What Claude Does |\n|--------|---------|------------------|\n| `executed` | Tasks were executed | Continues iterating |\n| `waiting` | Breakpoint or sleep gate | Asks for your approval |\n| `completed` | Run finished successfully | Reports final results |\n| `failed` | Run encountered an error | Reports error and recovery options |\n\n### Quality Gates\n\nBabysitter uses quality gates to ensure work meets standards:\n\n- **Quality Score**: A 0-100 score based on test coverage, code quality, and requirements coverage\n- **Quality Threshold**: The minimum score required (default: 80%)\n- **Convergence**: Babysitter iterates until the threshold is met or max iterations reached\n\n---\n\n## Troubleshooting\n\n### Common First-Time Issues\n\n#### Issue: Skill not found\n\n**Symptom:** `/babysitter:call` doesn't work or skill not in `/skills` list\n\n**Solution:** Verify plugin installation:\n1. Check plugin directory exists: `ls ~/.claude/plugins/babysitter/`\n2. If you installed from source, re-run `npm run generate:plugins` before copying the bundle so the Claude Code files are generated from `plugins/babysitter-unified`\n3. Restart Claude Code\n4. Run `/skills` again\n\n#### Issue: \"jq: command not found\" during execution\n\n**Solution:** Install jq (required for hooks):\n\n```bash\n# macOS\nbrew install jq\n\n# Ubuntu/Debian\nsudo apt-get install jq\n\n# Windows\nchoco install jq\n```\n\n#### Issue: Run seems stuck\n\n**Cause:** Usually waiting for your approval at a breakpoint.\n\n**Solution:** Check if Claude is asking you a question. In interactive mode, breakpoints require your response to continue.\n\n#### Issue: Quality threshold never reached\n\n**Cause:** The implementation may need significant changes to meet the threshold.\n\n**Solutions:**\n- Lower the quality threshold: \"targeting 70% quality\"\n- Increase max iterations: \"with up to 10 iterations\"\n- Simplify requirements\n\n#### Issue: Permission denied errors\n\n**Solution:** Make hook scripts executable:\n\n```bash\nchmod +x plugins/babysitter-unified/hooks/*.sh\n```\n\n### Health Check\n\nRun the health check script to diagnose issues:\n\n```bash\n```\n\n**Health status meanings:**\n\n| Status | Exit Code | Meaning |\n|--------|-----------|---------|\n| HEALTHY | 0 | All checks passed |\n| DEGRADED | 2 | Warnings present but functional |\n| UNHEALTHY | 1 | Critical failures detected |\n\n---\n\n## Next Steps\n\nNow that you understand how to use Babysitter, explore these resources:\n\n### Choose Your Methodology\n\nSee the [Process Selection Guide](https://github.com/a5c-ai/babysitter/blob/main/docs/reference/PROCESS_SELECTION.md) for a guide on choosing the right methodology:\n\n| Methodology | Best For |\n|-------------|----------|\n| **GSD** | Quick prototypes, getting things done fast |\n| **TDD Quality Convergence** | Quality-focused development with tests |\n| **Spec-Kit** | Projects with formal specifications |\n| **Scrum/Agile** | Sprint-based team workflows |\n| **BDD** | Behavior-driven development |\n\n### Example Invocations\n\n```\n# Quick prototype\n/babysitter:call use GSD to create a CLI tool for converting CSV to JSON\n\n# Quality-focused\n/babysitter:call implement user authentication with TDD, 85% quality threshold\n\n# Spec-driven\n/babysitter:call use Spec-Kit to implement the API defined in openapi.yaml\n\n# Feature development\n/babysitter:call add dark mode support to the React app using iterative development\n```\n\n### Advanced Topics\n\n- **[Unified Plugin Source](https://github.com/a5c-ai/babysitter/tree/main/plugins/babysitter-unified)** - Maintained plugin source tree\n- **[Advanced Patterns](https://github.com/a5c-ai/babysitter/blob/main/docs/reference/ADVANCED_PATTERNS.md)** - Agent tasks, parallel execution\n- **[Troubleshooting Guide](https://github.com/a5c-ai/babysitter/blob/main/docs/reference/TROUBLESHOOTING.md)** - Comprehensive problem-solving\n\n### Utility Scripts\n\n---\n\n**Need help?** Just ask Claude: \"Use the babysitter skill to help me with...\" and describe what you want to build.\n",
"documents": []
},
"outgoingEdges": [],
"incomingEdges": []
}