1tsukim/dotfiles
Personal dotfiles for macOS — zsh / git / tmux / Claude Code skill configs
ARCHETYPE
YOLO Cowboy
Permissions? Never heard of 'em. Ships at the speed of `--dangerously-skip-permissions`.
Copy this rig
# review before running: this installs third-party code
$ npx degit 1tsukim/dotfiles/claude/.claude ./rig-dotfiles # inspect, then merge into .claude/ MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ curl -fsSL --create-dirs -o .claude/agents/ai-writing-auditor.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/ai-writing-auditor.md $ curl -fsSL --create-dirs -o .claude/agents/ai-writing-auditor.vocab.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/ai-writing-auditor.vocab.md $ curl -fsSL --create-dirs -o .claude/agents/code-review-specialist.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/code-review-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/codebase-analyzer.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/codebase-analyzer.md $ curl -fsSL --create-dirs -o .claude/agents/data-scientist.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/data-scientist.md $ curl -fsSL --create-dirs -o .claude/agents/debug-specialist.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/debug-specialist.md $ curl -fsSL --create-dirs -o .claude/agents/ml-pipeline-optimizer.md https://raw.githubusercontent.com/1tsukim/dotfiles/main/claude/.claude/agents/ml-pipeline-optimizer.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"permissions": {
"deny": [
"Bash(sudo:*)",
"Bash(rm -rf /)",
"Bash(rm -rf ~)",
"Bash(rm -rf ~/*)",
"Bash(rm -rf /*)",
"Read(.env.*)",
"Read(id_rsa)",
"Read(id_ed25519)",
"Read(**/*token*)",
"Read(**/*key*)",
"Read(**/.aws/**)",
"Read(**/.ssh/**)",
"Read(**/credentials)",
"Read(**/credentials.*)",
"Read(**/*credential*)",
"Read(**/*secret*)",
"Read(**/*password*)",
"Edit(.env*)",
"Edit(**/secrets/**)",
"Bash(npm uninstall:*)",
"Bash(npm remove:*)",
"Bash(xargs rm:*)",
"Bash(xargs trash:*)",
"Bash(xargs sh:*)",
"Bash(xargs bash:*)",
"Bash(xargs sudo:*)",
"Bash(*xargs rm*)",
"Bash(*xargs *rm *)"
]
},
"hooks": {
"PreToolUse": [
{
"matcher": "Bash",
"hooks": [
{
"type": "command",
"command": "~/.claude/hooks/block-dangerous-git.sh"
},
{
"type": "command",
"command": "~/.claude/hooks/block-dangerous-aws.sh"
},
{
"type": "command",
"command": "~/.claude/hooks/block-dangerous-misc.sh"
},
{
"type": "command",
"command": "~/.claude/hooks/block-sensitive-bash.sh"
},
{
"type": "command",
"command": "~/.claude/hooks/check-s3-cp-overwrite.sh"
},
{
"type": "command",
"command": "jq -r '\"[\\(now | strftime(\"%Y-%m-%d %H:%M:%S\"))] \\(.tool_input.command)\"' >> ~/.claude/bash-history.log 2>/dev/null || true"
}
]
},
{
"matcher": "Write|Edit",
"hooks": [
{
"type": "command",
"command": "~/.claude/hooks/block-sensitive-files.sh"
}
]
}
]
}
} Subagents (7)
| ai-writing-auditor model: opus | Detect and remove AI-generated writing patterns (AI-isms) from Japanese business documents. **Primary use case: post-process Claude's own output before delivery** — analysis reports, proposal docs, me |
| ai-writing-auditor.vocab | — |
| code-review-specialist model: opus | Use this agent when you have written or modified code and need a comprehensive multi-axis quality review (quality + security + performance + maintainability) in an isolated context. Best for substanti |
| codebase-analyzer model: sonnet | Use this agent for exploratory comprehension of an unfamiliar codebase — examining the structure, entry points, and architecture, and returning a synthesized overview. Best when you need a mental mode |
| data-scientist model: sonnet | Use this agent for hypothesis-driven exploratory data analysis, statistical hypothesis testing, causal inference, and translating findings into reports. Best for Notebook-style analytics work where th |
| debug-specialist model: opus | Use this agent when a bug requires multi-file investigation, hypothesis testing, or actual tool execution (grep / edit / run commands) beyond a quick read. The agent runs in an isolated context and ex |
| ml-pipeline-optimizer model: sonnet | Use this agent when the user is explicitly working on an ML pipeline — implementing training/inference code, performing feature engineering, or seeking to improve an existing model's accuracy. Trigger |
Hooks (12)
| event | matcher | runs |
|---|---|---|
| Notification | * | ~/.claude/hooks/notify.sh notification |
| Stop | * | ~/.claude/hooks/notify.sh stop |
| SubagentStop | * | ~/.claude/hooks/notify.sh subagent-stop |
| PreToolUse | Bash | ~/.claude/hooks/block-dangerous-git.sh |
| PreToolUse | Bash | ~/.claude/hooks/block-dangerous-aws.sh |
| PreToolUse | Bash | ~/.claude/hooks/block-dangerous-misc.sh |
| PreToolUse | Bash | ~/.claude/hooks/block-sensitive-bash.sh |
| PreToolUse | Bash | ~/.claude/hooks/check-s3-cp-overwrite.sh |
| PreToolUse | Bash | jq -r '"[\(now | strftime("%Y-%m-%d %H:%M:%S"))] \(.tool_input.command)"' >> ~/.claude/bash-history.log 2>/dev/null || true |
| PreToolUse | Write|Edit | ~/.claude/hooks/block-sensitive-files.sh |
| PostToolUse | DISABLED | ~/.claude/hooks/auto-format-python.sh |
| SessionStart | * | { git status --short 2>/dev/null || echo '(not a git repo)'; } | jq -Rs '{hookSpecificOutput:{hookEventName:"SessionStart",additionalContext:("git status:\n" + .)}}' 2>/dev/null || true |
Slash commands (3)
/fix-github-issue/gemini-search/serena
Permissions
deny (28)
Bash(sudo:*)
Bash(rm -rf /)
Bash(rm -rf ~)
Bash(rm -rf ~/*)
Bash(rm -rf /*)
Read(.env.*)
Read(id_rsa)
Read(id_ed25519)
Read(**/*token*)
Read(**/*key*)
Read(**/.aws/**)
Read(**/.ssh/**)
Read(**/credentials)
Read(**/credentials.*)
Read(**/*credential*)
Read(**/*secret*)
Read(**/*password*)
Edit(.env*)
Edit(**/secrets/**)
Bash(npm uninstall:*)
Bash(npm remove:*)
Bash(xargs rm:*)
Bash(xargs trash:*)
Bash(xargs sh:*)
Bash(xargs bash:*)
Bash(xargs sudo:*)
Bash(*xargs rm*)
Bash(*xargs *rm *)
ask (0)
—
allow (57)
WebSearch
WebFetch
Skill
Read
Read(**)
Glob
Grep
LSP
Bash(ls:*)
Bash(cat:*)
Bash(head:*)
Bash(tail:*)
Bash(find:*)
Bash(which:*)
Bash(pwd:*)
Bash(tree:*)
Bash(wc:*)
Bash(file:*)
Bash(stat:*)
Bash(echo:*)
Bash(gh:*)
Bash(aws:*)
Edit(.tmp/**)
mcp__context7__resolve-library-id
mcp__context7__get-library-docs
mcp__serena
mcp__playwright__browser_take_screenshot
mcp__playwright__browser_wait_for
Bash(rg:*)
Bash(fd:*)
Bash(bat:*)
Bash(eza:*)
Bash(jq:*)
Bash(xargs:*)
Bash(uv:*)
Bash(grep:*)
Bash(mkdir:*)
Bash(mv:*)
Bash(cp:*)
Bash(touch:*)
Bash(python:*)
Bash(python3:*)
Bash(trash:*)
Bash(chmod:*)
Bash(ln:*)
Bash(marp:*)
Bash(git status:*)
Bash(git diff:*)
Bash(git log:*)
Bash(git show:*)
Bash(git blame:*)
Bash(git ls-files:*)
Bash(git rev-parse:*)
Bash(git stash list)
Bash(git branch:*)
Bash(git remote:*)
Bash(git config --get*)
Similar rigs
humanlayer/humanlayer
The best way to get AI coding agents to solve hard problems in complex codebases.
Orchestrator 2.7k tok ·
nexusweblabs/claude-bedrock-setup
CLI tool to configure Claude Code for AWS Bedrock
Orchestrator 1.2k tok ·
jeffersonBastos/precis
A comprehensive template for Claude Code projects with pre-configured commands, agents, and workflows.
Orchestrator 982 tok ·
medusajs/medusa
The world's most flexible commerce platform for agents and developers
Pragmatist 4.4k tok ·