Policy enforcement for every agent action.
RunAgents evaluates each tool call in context before it reaches downstream systems. Teams can allow safe operations, require approval for risky actions, and deny dangerous writes by default.
- Allow, deny, and approval-required behavior in one policy model
- Safer agent access to CRMs, ticketing systems, collaboration tools, and internal APIs
- A clearer security review story for production adoption
Start here: Try Alpha, Docs, API Reference, GitHub.