agent-boundary-adversary
Tests whether an isolated AI sandbox actually blocks file access, credentials, network, and commands before you use it in production.
Installation
Paste this into Claude Code, Cursor, or any agent that can run commands.
What this skill does
What it does:
- Tests whether a sandboxed computer environment properly blocks dangerous actions
- Tries to access files, passwords, network, and commands that should be blocked
- Records what happens when each test runs
- Confirms that the sandbox is working correctly before using it for real work
When to use it:
- You set up a new isolated computer space and need to check it is safe
- You changed the security settings and want to make sure they still work
- You need proof that a sandbox blocks what it is supposed to block
- Before letting an AI system run in a restricted environment for the first time
The full skill text could not be loaded right now. It installs fine with the command above, or read it at the source.
Mirrored from the author's public source. Install counts from the open skills registry.