my-security-review-checklistlisted

Pre-merge security gate for agent tooling - skills, subagents, slash commands, hooks, shell and sync scripts, dotfiles, settings.json, and plugin/MCP trust. Also scans shared tooling for user-specific absolute paths, device overfitting and hardcoded personal constants, and can fix them. Scans any file an agent reads as instructions for text hidden from the human reviewer (invisible Unicode, tag smuggling, zero-width characters, bidi overrides), so use it whenever asked whether a skill, prompt, rules file or pasted content has hidden instructions in it. MUST be used before committing any change to ~/.agents or ~/.claude, before keeping a new or changed hook, before trusting a plugin or MCP server, and for any automation that consumes untrusted agent/LLM/web output.
anton-winter-arch/dotagents · ★ 1 · AI & Automation · score 72

Install: claude install-skill anton-winter-arch/dotagents

# My Security Review Checklist ## Overview A tight, runnable security gate for the kind of code in this ecosystem: **agent skills, subagents, slash commands, hooks, shell and sync scripts, dotfiles, and plugin/MCP configuration.** This is not web-app security - for that, use `agent-skills:security-and-hardening`. This checklist exists because agent tooling has its own attack surface: arbitrary command execution via hooks, untrusted LLM/agent output flowing into actions, secrets sitting in config files, and destructive file operations on a synced source of truth. Run it **before merge, not after.** The review is the gate. ## When to Use - Before committing or merging any change to a **skill, subagent, command, or hook** - Before editing **`settings.json` / `settings.local.json`** (permissions, env, hooks) - Before changing a **shell or sync script** (`sync-skills.sh`, anything with `mv`/`rm`/`ln`) - Before **installing, updating, or trusting a plugin or MCP server** - Whenever code will **consume untrusted input** - agent inbox messages, web/MCP responses, file contents, command output ## The Checklist Findings are labelled **Critical** (blocks merge), **Important** (fix before merge), or **Suggestion** (optional). Treat anything touching secrets, command execution, or destructive file ops as Critical until proven otherwise. ### 1. Secrets & dotfiles - [ ] No secrets (API keys, tokens, passwords) in any committed file - skills, hooks, scripts, configs - [ ] `.gitignor