gate.cat Answers How it works pip install gate-cat

answers · agent safety

answers

Plain answers to what people ask right after an AI agent does something it can't take back. Every answer ends the same way: a deterministic, fail-closed veto that stops the irreversible command before it runs — free, Apache 2.0, one line to install.

My AI agent almost ran rm -rf — how do I stop it?
Why prompts don't enforce, and what a fail-closed veto does instead.
How do I secure an MCP server against destructive tool calls?
Put the veto in front of the tools, not inside the model.
Claude Code deleted my files — recovery & prevention
What to try first for recovery, then how to stop the next one.
How do I get an audit trail of what my AI agent did?
A free local log of every irreversible action, and what it is (and isn't).
How do I enforce guardrails on AI agents across my team?
Free per machine; a fleet record, alerts and one policy set for the team.

the one-line version

gate.cat inspects every command an AI agent tries and blocks the irreversible ones — recursive delete, database drop, cloud teardown — before they execute. 38 default policies, fail-closed, zero telemetry.

$ pip install gate-cat

One email when there's a catch worth showing: a real irreversible command an agent tried, and how the wall stopped it. Opt-in, no spam, no telemetry.

✓ you're on the list — first catch lands soon.