Skip to content

Latest commit

 

History

History
executable file
·
216 lines (171 loc) · 10.8 KB

File metadata and controls

executable file
·
216 lines (171 loc) · 10.8 KB

bertini36/dotfiles 

My personal Mac setup and configurations

🚀 Setup

  • Download code:

    git clone https://github.com/bertini36/dotfiles.git ~/.dotfiles/
  • Brew packages installation:

    brew bundle --file=mac/Brewfile
    Package Description
    bat cat with syntax highlighting
    eza Modern ls replacement
    fzf Fuzzy finder for the terminal
    gh GitHub CLI
    pre-commit Git hook manager
    graphviz Graph visualization tools
    jq JSON processor
    libmagic File type detection library
    gotop Terminal system monitor
    gemini-cli Google Gemini AI CLI
    copilot-cli GitHub Copilot CLI (cask)
    mole macOS disk space cleaner and system optimizer
    postgresql@18 PostgreSQL database
    pyenv Python version manager
    uv Fast Python package manager
    python@3.14 Python interpreter
    tldr Simplified man pages with practical examples
    karabiner-elements Keyboard remapper (cask)
    fd Fast find replacement
    ripgrep Fast grep replacement
    semgrep Static analysis (SAST) scanner
    gitleaks Secret detection in git commits
    nvm Node version manager
    pnpm Fast Node package manager
    claude Anthropic Claude desktop app (cask)
    claude-code Anthropic Claude CLI (cask)
    rtk CLI proxy that reduces LLM token consumption by 60-90%
    handy Speech-to-text utility
  • Extra configuration (not available through Brew):

    bash mac/config_extras.sh
    Config Description
    gitleaks hook Global git pre-commit hook for secret detection
  • Add fonts (fonts/) to Font Book

  • Configure Karabiner

    • Change Caps Lock to CMD + CTL + Option + Shift
    • Map F4 to CMD + Space (Raycast)
  • Install Oh My ZSH

    • Link shell/.zshrc to ~/.zshrc: ln ~/.dotfiles/shell/.zshrc ~/.zshrc

    • Install plugins

      git clone https://github.com/zsh-users/zsh-syntax-highlighting.git ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/plugins/zsh-syntax-highlighting
      git clone https://github.com/zsh-users/zsh-autosuggestions ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/plugins/zsh-autosuggestions
      git clone https://github.com/agkozak/zsh-z ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/plugins/zsh-z
  • Install Raycast

    • Disable Spotlight shortcut to enable Raycast one (System Preferences -> Keyboard -> Shortcuts -> Spotlight -> Uncheck Show Spotlight search)
    • Configure shortcuts following keymap.md
  • Install Iterm2

  • Install Docker

  • Install Jetbrains Toolbox and Pycharm

  • Install Visual Studio Code

  • Enable auto-focus: defaults write com.apple.Terminal FocusFollowsMouse -bool true

  • Link the rest of configuration files

    ln -s ~/.dotfiles/git/.gitignore_global ~/.gitignore_global
    git config --global core.excludesfile ~/.gitignore_global
    
    ln -s ~/.dotfiles/editors/vim/.vimrc ~/.vimrc
    
    ln -s ~/.dotfiles/editors/claude/settings.json ~/.claude/settings.json
    ln -s ~/.dotfiles/editors/claude/statusline-command.sh ~/.claude/statusline-command.sh
    ln -s ~/.dotfiles/editors/claude/CLAUDE.md ~/.claude/CLAUDE.md
    ln -s ~/.dotfiles/editors/claude/skills ~/.claude/skills
    ln -s ~/.dotfiles/editors/claude/rules ~/.claude/rules
    ln -s ~/.dotfiles/editors/claude/agents ~/.claude/agents
    ln -s ~/.dotfiles/editors/claude/commands ~/.claude/commands

🧠 Claude Configuration

All Claude Code configuration lives under editors/claude/ and is symlinked into ~/.claude/.

Workflow

See editors/claude/workflow.md for detailed documentation of the development workflow using Claude, Superpowers, and the custom skills, agents, rules, and commands defined in this repository.

Skills

Reusable AI agent skills that Claude invokes autonomously when a task matches their description.

Skill Description
create-pull-request Create a GitHub PR following project conventions using gh CLI
ddd-patterns DDD entities, aggregate roots, value objects, repositories, domain services, and specifications
django-patterns Django architecture, REST APIs with Pydantic, ORM best practices, caching, and signals
langchain-architecture LangChain 1.x and LangGraph for agents, memory, and tool integration
production-code-audit Deep-scan a codebase and transform it to production-grade quality
python-code-style Python type safety, generics, protocols, and advanced type annotations
writing-clearly-and-concisely Clear prose for docs, commits, error messages, and UI text

Evals

Each skill has an evals/evals.json file that defines test cases to measure skill effectiveness. To run the evals paste the following command your AI agent prompt.

  1. Read the eval definitions in editors/claude/skills/<skill>/evals/evals.json
  2. Generate outputs - run each eval prompt twice per skill (once with the skill loaded, once without) and save the results to editors/claude/skills-workspace/iteration-1/<eval-id>/with_skill/outputs/ and without_skill/outputs/
  3. Create eval_metadata.json - record the assertions from each eval's expectations array alongside references to the output files
  4. Compare outputs in with_skill/outputs/ vs without_skill/outputs/
  5. Verify each assertion from eval_metadata.json against the corresponding output

Agents

Specialized subagents that run in isolated context windows with restricted tools.

Agent Description
code-reviewer Read-only production code audit with A-F graded report (architecture, security, performance, quality, testing)
security-reviewer OWASP Top 10 and Django-specific security vulnerability scanner
evaluator Quality gate that scores implementation plans on 7 criteria with GO/NO-GO verdict

Rules

Path-scoped rules that load automatically only when working on matching files.

Rule Scope
python **/*.py - Python 3.12+ conventions, ruff, uv, naming, imports
django Django files (views, models, urls, admin, etc.)
tests Test files - no comments, self-explanatory naming
langchain LangChain/LangGraph files

Commands

Custom slash commands for common workflows.

Command Usage
/review Review current branch changes for quality and security
/fix-issue <number> Fetch a GitHub issue and implement the fix
/audit Run full production audit with both agents

🔌 Claude Plugins

Install the following MCP server plugins in Claude:

Plugin Description
superpowers Spec driven development (SDD) based on brainstorming, planning, subagent-driven execution, TDD, and code review skills
context7 Up-to-date documentation and code examples for any library
claude-mem Persistent cross-session memory database with smart search and timeline reports (localhost:37777)
sentry-skills Sentry engineering skills: PR writing, code review, Django patterns, security review, and more
caveman Caveman-speak mode that cuts ~75% of output tokens while keeping technical accuracy
notion Read and manage Notion pages and databases
figma Read Figma designs and generate code from them
datadog-mcp Datadog observability: logs, metrics, traces, incidents, monitors, and dashboards

Install Datadog MCP:

claude mcp add --transport http datadog-mcp https://mcp.datadoghq.eu/api/unstable/mcp-server/mcp

Install Caveman:

claude plugin marketplace add JuliusBrussee/caveman
claude plugin install caveman@caveman

Activate rtk:

rtk init -g

Built with ❤️ from Mallorca