AgentGuard
Documentation
Two doorways: one for users getting AgentGuard running, one for developers wiring the API into their own tools.
Quickstart for users
Install the AgentGuard CLI, hook it into your agent, save your first policy, and verify that risky commands actually get caught. ~5 minutes.
Open quickstartAPI reference
Full request / response schema for every endpoint — scan, runtime evaluate, events ingest, approvals, sessions, status. For when you're building on top.
Open API reference