🚀 Best of: Claude Code – Agents, Hooks & Git Magic

The most essential learnings from community projects, blogs & docs – for a robust, reproducible AI dev pipeline. 👇

🧭 Agentic Workflows (Plan → Build → Verify)

• Multi-agent pipelines: o3 plans in detail, Sonnet builds, strict model verifies.
• Each task = its own commit; parallel with Git worktrees.

🧩 Sub-Agents & Meta-Agents

• Sub-agents as Markdown with YAML (Reviewer, Test Engineer, Docs, Security, Perf, Architect).
• Meta-agent generates new sub-agents incl. tooling & prompts – consistent format by design.

🛠️ Best Practices

• Phase 1: Read & Plan (structured steps).
• Phase 2: Implement & Validate (tests first, then commit). ✅

🪝 Hook System (8 Events)

UserPromptSubmit, Pre/PostToolUse, Notification, Stop, SubagentStop, PreCompact, SessionStart.

Use Cases:

• Prompt logging
• Security checks (block rm -rf / .env)
• Auto-formatting (Prettier/ruff/black)
• Desktop notifications, backups, audit logs

🧭 Deterministic Control Flow

Hooks return JSON (continue, feedback, modify_prompt) – validation & prompt correction before proceeding.

🌿 Git Automation

• GitButler: Branches & commits per session (Pre/PostToolUse + Stop).
• “Git Paranoia Mode”: Auto-formatter per stack, checkpoint commits, clean squash at the end.
• Pre-commit: Lint & tests; Post-edit: Format; Auto-docs for JS/TS.

🔔 More Automation

• Apidog configs (matcher-based)
• ntfy.sh push & TTS summaries
• Regex security layer in pre_tool_use.py

📊 Observability

Multi-agent dashboard (HTTP → SQLite → WebSockets), real-time view, replay, token tracking & burn-rate alerts.

🔗 Slash Commands & MCP

/code-review, /refactor, /update-docs … + MCP for GitHub/Jira/APIs (PR reviews, tickets, parallel research).

🛡️ Validation & Security

Exit codes: 0=go, 1=warn, 2=block, 3=stop. Protect sensitive files, compliance checks before commits.

Conclusion

Plan, automate, verify – and Git integrations + hooks turn Claude Code into a reliable pair programmer. 🧠⚙️