Omamori — AI skill for Claude Code
AI Agent's Omamori — protect your system from dangerous commands executed via AI CLI tools.
How to install Omamori
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open yottayoshida/omamori and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Omamori does
AI Agent's Omamori — protect your system from dangerous commands executed via AI CLI tools.
Alternatives in AI
- System Prompts Leaks — Extracted system prompts from ChatGPT (GPT-5.4, GPT-5.3, Codex), Claude (Opus 4.6, Sonnet 4.6, Claude Code), G 38.6k ★
- AIHawk — Open-source AI browser agent: describe any task in plain language and it autonomously browses, clicks, types a 30.3k ★
- Qwen Code — A command-line AI workflow tool adapted from Gemini CLI, optimized for Qwen3-Coder models with enhanced parser 20.8k ★
README
omamori
[](https://github.com/yottayoshida/omamori/actions/workflows/ci.yml) [](https://crates.io/crates/omamori) [](https://github.com/yottayoshida/homebrew-tap) [](LICENSE-MIT)
Deterministic semantic guard for AI CLI tools. Blocks covered destructive commands and self-disablement attempts, with tamper-evident audit trails.
Fast local checks — no model calls, no daemon, no network dependency. **macOS only.** Commands you type in a plain terminal pass through untouched; omamori acts only when an AI tool environment is detected. It is not a sandbox: run it alongside one ([how the two fit together](docs/reference-architecture.md)).
Since **1.0**, three surfaces are frozen until a major version: which command classes are blocked or redirected, the CLI's subcommands and documented exit codes, and the audit chain's verifiability across upgrades. The `config.toml` schema and the Rust library API are not frozen. The full statement is [docs/CONTRACT.md](docs/CONTRACT.md).
Demo

A Claude Code session. Codex CLI and Cursor get the same `block` / `log-only` / `trash` behavior when detected.
Quick start
brew install yottayoshida/tap/omamori
omamori setup
`setup` installs the shims and hooks, adds `$HOME/.omamori/shim` to your shell profile, and runs `omamori doctor`. It is safe to re-run after upgrades. Preview with `--dry-run`; use `--non-interactive` in CI and scripts.
To see what omamori would do with a command, without running it:
omamori explain -- rm -rf src/
What it does
- Blocks destructive commands before they run.
rm -rfgoes to the macOS Trash,git reset --hardstashes first, andgit push --force,git clean -f,chmod 777, `find -delete
Related Skills
AI Guard Coding Agents
A CLI that runs AI coding agent actions through Datadog AI Guard before they are executed.
MCP Unifi Applications
UniFi MCP server that makes the official UniFi API documentation (Network, Protect, Site Manager, InnerSpace)
Fireraven Agent Hooks
FireGuard hooks for AI agents: protect Cursor, Claude Code, Windsurf, Devin, and Copilot Studio from secret le
Sealgate
Protect sensitive prompt spans with trusted LLM detection and local authenticated encryption before using Clau
AuRO Whisky macOS Setup
AI-executed skill that installs uaRO (Ragnarok Online private server) on Apple Silicon Macs via Whisky/Wine —
Pythtb Skill
AI-agent skill, verified toolkit, fully executed notebooks and an undergraduate course for PythTB 2.0 - tight-
Related Agents
Codex Analyst
Token-heavy READ-ONLY analysis, codebase surveys, audits and inventories, executed by the Codex CLI rather tha
Aegis
You are a specialized security agent. Your job is to identify vulnerabilities, analyze security risks, and rec
Review Test Crash Safety
Reviews test code for crash safety testing quality and production assert statements: whether crash/recovery sc