Two ready-to-use templates for Claude Code projects.
Copy this into any new project to get a working Claude Code setup instantly.
Directory layout:
project-boilerplate/
├── CLAUDE.md # Engineering rules
├── .mcp.json # MCP server defaults (playwright)
├── pyproject.toml # UV-based Python project
├── env.sample.txt # Environment variable template
├── .gitignore # Python, logs, env, IDE, Claude runtime
├── .gitattributes # Line endings, diff drivers
├── apps/ # Application code
├── ai_docs/ # Fetched documentation cache
│ └── README.md # URL list for /load-ai-docs
├── specs/ # Implementation plans (/plan output)
├── app_review/ # Code review reports (/review output)
├── app_fix_reports/ # Fix reports (/fix output)
├── scripts/ # Utility scripts
├── tests/ # Test files
├── logs/ # Runtime logs (gitignored)
└── .claude/
├── settings.json # Status line, hooks, permissions
├── hooks/
│ ├── pre_tool_use.py # Damage control
│ ├── post_tool_use.py # Logging
│ └── user_prompt_submit.py # Prompt tracking
├── commands/
│ ├── plan.md # Create implementation plan -> specs/
│ ├── build.md # Implement plan top-to-bottom
│ ├── review.md # Risk-tiered code review -> app_review/
│ ├── fix.md # Fix review issues -> app_fix_reports/
│ ├── question.md # Read-only project Q&A
│ ├── load-ai-docs.md # Fetch docs from URLs -> ai_docs/
│ └── meta-prompt.md # Create new slash commands
├── agents/
│ └── docs-scraper.md # URL -> markdown scraping agent
├── output-styles/ # 11 output formatting styles
└── status_lines/ # Status line scripts (v1 + v4)
ADW workflow: /plan -> /build -> /review -> /fix
Usage:
cp -r project-boilerplate/ ~/my-new-project/
cd ~/my-new-project
uv syncOnboarding toolkit for new engineers. Combines deterministic hooks with agentic commands and validation scripting.
Hooks & Validators:
session_start.py— Loads.envinto Claude Code sessionsetup_init.py— Runsuv syncand project-specific init steps onclaude --initvalidators/bash_validator.py— Blocks dangerous bash patterns (PreToolUse)validators/ruff_validator.py— Lints Python on Write (PostToolUse)validators/validate_new_file.py— Verifies file creation (Stop hook)validators/validate_file_contains.py— Verifies file content requirements (Stop hook)
Commands (13 total, - normalized):
| Command | Purpose |
|---|---|
/all-tools |
List all available tools with signatures |
/biweekly |
Generate HTML progress report from GitHub issues/PRs |
/install |
Deterministic setup verification |
/load-ai-docs |
Fetch documentation URLs into local markdown |
/maintenance |
Dependency updates and health checks |
/meta-prompt |
Meta-prompt that creates new slash commands |
/onboard-candidate |
Create assessment repo + collaborator + HTML email |
/question |
Read-only project Q&A |
/reverse-goal-setting |
Skill-first goal setting framework |
/review-candidate |
Weighted assessment evaluation with comparison |
/semantic-commit |
Group changes semantically, create issue/branch/PR |
/sentient |
Test damage control guards |
/spawn-project |
Deterministic port allocation for orchestrator instances |
Justfile recipes:
just help # List all recipes
just check # Verify dependencies (claude, uv, gh, git, ruff)
just cldi # Deterministic init (hook only, CI-friendly)
just cldii # Agentic init (hook + analysis)
just cldit # Interactive init (hook + agent + questions)
just tools # /all-tools
just question "?" # /question
just metaprompt # /meta-prompt
just commit # /semantic-commit
just biweekly # /biweekly
just load-docs # /load-ai-docs
just spawn proj # /spawn-project- Claude Code CLI
- Astral UV for Python
- just command runner
- gh GitHub CLI