PRB's .claude directory.
Get up and running in 3 steps:
# 1. Clone to ~/.claude
git clone [email protected]:PaulRBerg/dot-claude.git ~/.claude
cd ~/.claude
# 2. Install dependencies
just install
# 3. Try it out
ccc # Make your first commit with the claude wrapperSee Installation for detailed setup and Configuration for customization.
- Node.js - For Husky/lint-staged automation (
npm install) - Just - Command runner for build scripts (
brew install just) - Python 3.13+ and uv - Python package and project manager
Clone or copy this repository to ~/.claude:
git clone [email protected]:PaulRBerg/dot-claude.git ~/.claude
cd ~/.claudejust install # Node deps, Python deps, and CLI utilitiesjust full-check # Run all code checks
just test-hooks # Run hook tests
claude # Run ClaudeModular architecture: All JSONC files in settings/* automatically merge into settings.json on commit via Husky +
lint-staged.
Editing: Modify settings/**/*.jsonc files (never edit settings.json directly). Changes auto-merge on commit, or
run just merge-settings manually.
Global instructions live in CLAUDE.md. Sections cover communication style, code preferences, scope preservation, git
safety, and shell escaping rules.
Add flags at the end of prompts to trigger behaviors (-s for subagents, -c for auto-commit, -t for testing, -d
for debug, -n to skip linting). Flags are composable: implement API -s -t -c.
Run common tasks like just full-check (all code checks), just merge-settings (merge JSONC files), or just test
(run tests). See justfile for all commands.
Slash commands in commands/*.md handle GitHub workflows, releases, and task management:
/create-pr- Create PRs with semantic analysis/create-issue- Create GitHub issues/update-pr- Update existing PRs/create-discussion- Start GitHub discussions/update-deps- Update Node.js dependencies/md-docs:update-readme- Sync README with codebase/md-docs:update-agents- Update AI context files
Activatable skills live in ~/.agents/skills. Examples: typescript,
gh-cli, code-review.
Specialized subagents in agents/: code review, debugging, tool discovery, docs finding. Invoke via -s flag or Task
tool.
Two servers extend Claude's capabilities (configured in .mcp.json):
- context7 - Library documentation and code examples
- sequential-thinking - Chain-of-thought reasoning
Enable/disable in settings/permissions/mcp.jsonc.
Event-driven automation for Claude Code events. See hooks/README.md for details.
Key hooks:
- ai-flags - Parse flags from prompts to trigger behaviors
- ai-notify - Desktop notifications for events (optional)
- log_prompts.py - Log conversations to zk notebook (optional)
Note
The claude wrapper runs Claude with --dangerously-skip-permissions and auto-loads MCP servers from .mcp.json.
See this issue.
Optional shell utilities in utils.sh:
ccc [args]- Streamlined commits via/commit(defaults to--all)cccp- Commit and push (for feature branches)ccbump [args]- Quick release bumping via/bump-releaseclaude [args]- Enhanced CLI wrapper with MCP auto-loading
Source in your shell config:
# In ~/.zshrc
source ~/.claude/utils.shThis project is licensed under MIT - see the LICENSE.md file for details.