I wanted to leave Claude Code running without trusting a green checkmark

A developer known as CraniusMaximus built the Autonomy Pack, a set of hooks and operator tooling for Claude Code, to prevent unattended runs from fabricating success, skipping checks, or incurring excessive costs. The pack, priced at $29, pairs with a session-tax plugin for cost reporting, and the developer acknowledges it does not make agents safe but aims to reduce failure modes in overnight sessions.

Unattended Claude Code runs are tempting. They are also how you wake up to a fabricated "done" and a burned token budget. I built the Autonomy Pack because I wanted guardrails and hooks I could actually walk away from. Not vibes. Hooks that catch the failure modes I kept hitting when I left a session overnight. Hooks and operator tooling so an unattended run is less likely to invent success, skip checks, or quietly do the expensive wrong thing. Pair it with session-tax if you also care what the session costs before it starts. I will not pretend this makes agents safe. It makes my unattended runs less stupid. Your mileage will depend on how you work. GitHub README trusted link for forums that filter Pages URLs : github.com/CraniusMaximus/claude-code-autonomy-pack https://github.com/CraniusMaximus/claude-code-autonomy-pack Paid pack is $29. Bundle with session-tax is $39. Free session-tax plugin is listed there too if you only want the cost report. If you already leave Claude Code unattended, what actually broke on you? Fabricated completions, hooks that never fired, runaway context — I am collecting the failure modes, not a feature wishlist.