claude-code-mastery
Dangerous skill for installing and managing Claude Code, subagents, and optional claude-mem. It runs remote installers like curl -fsSL https://claude.ai/install.sh | sh, executes local services, requests ANTHROPIC_API_KEY, and installs/modifies files under ~/.claude/agents/.
Claude Code Mastery 🧑💻
A comprehensive skill for mastering Claude Code with setup scripts, dev team subagents, and automated maintenance.
Why This Skill?
Claude Code is powerful on its own. This skill adds:
- Specialized subagents — Route tasks to experts (frontend, backend, PM, etc.)
- Structured setup — Scripts that handle installation correctly
- Self-improvement — Heartbeat tasks that keep you learning
- Best practices — Docs on context management, workflows, pro tips
Quick Install
cd ~/clawd/skills/claude-code-mastery/scripts
# Run setup scripts in order
./01-check-dependencies.sh
./02-install-claude-code.sh
./03-first-time-auth.sh
./04-install-subagents.sh # Starter pack (3 agents) - recommended
./04-install-subagents.sh --full-team # All 11 agents
./05-setup-claude-mem.sh # Optional - prompts y/N (default: No)
What's Included
🤖 Dev Team Subagents
Starter Pack (default):
| Agent | Purpose |
|---|---|
| senior-dev | Architecture, complex code, code review |
| project-manager | Task breakdown, timelines, dependencies |
| junior-dev | Quick fixes, simple tasks (fast & cheap) |
Full Team (--full-team):
| Agent | Purpose |
|---|---|
| frontend-dev | React, UI, CSS, client-side |
| backend-dev | APIs, databases, server-side |
| ai-engineer | LLM apps, RAG, prompts, agents |
| ml-engineer | ML models, training, MLOps |
| data-scientist | SQL, analysis, statistics |
| data-engineer | Pipelines, ETL, data infrastructure |
| product-manager | Requirements, user stories, prioritization |
| devops | CI/CD, Docker, K8s, infrastructure, automation |
Each agent includes a "Learn More" section with curated links to official docs, tutorials, and best practices.
📜 Scripts
| Script | Purpose |
|---|---|
01-check-dependencies.sh | Verify system requirements |
02-install-claude-code.sh | Install Claude Code CLI |
03-first-time-auth.sh | Authenticate (browser or API key) |
04-install-subagents.sh | Install subagents (--minimal or --full-team) |
05-setup-claude-mem.sh | Persistent memory (optional, prompts y/N) |
06-diagnostics.sh | Health check and status report |
07-weekly-improvement-cron.sh | Generate improvement report |
08-troubleshoot.sh | Comprehensive troubleshooting |
uninstall.sh | Clean removal of all components |
📚 Documentation
- SKILL.md - Complete usage guide
- docs/best-practices.md - Context management, verification tips
- docs/commands.md - CLI and slash command reference
- docs/workflows.md - Real-world workflow examples
- docs/tips-and-tricks.md - 30 pro tips from heavy users
- docs/troubleshooting.md - Common issues and fixes
⚙️ Configuration
Edit config.sh to customize:
VALID_MODELS— Add new models as Anthropic releases themHEARTBEAT_DIAGNOSTICS— Enable/disable in heartbeatINSTALL_MODE— Default starter vs full team
Usage
After setup, use Claude Code with your dev team:
claude
> Use the senior-dev agent to review this code
> Have project-manager create a timeline for this feature
> Ask junior-dev to fix this typo
Self-Improvement
This skill is designed to improve over time. The heartbeat task includes:
- Weekly learning rotation through agent expertise areas
- Instructions to update skill files with new discoveries
- Broken link fixes and best practice updates
Each Clawdbot using this skill can contribute improvements back!
License
MIT License - See LICENSE
Contributing
See CONTRIBUTING.md for guidelines.