Concepts
Each page explains a part of the Straza model and opens with the diagram from How it works, with that part highlighted.
- Roles and accessWhat a role lets its holders do, and who makes and assigns roles.
- ObjectsHow a role can be limited to some of a server's objects, on any server that declares its objects.
- PoliciesHow a rule matches a tool call, how Straza picks the rule that decides, and what each outcome does.
- ApprovalsWho decides a call that a rule holds, where they decide it, how long it waits, and what the agent meets in a hook and at the gateway.
- DraftsWhy each change to servers, roles and policy waits as a checked draft until a person publishes it, and who may propose one.
- People, agents and sponsorsWho a governed action belongs to, how Straza tells a person from an AI agent, and which person is accountable for an agent.
- Machines and sessionsHow a person's machine enrolls with a key that never leaves it, how each session gets a short-lived token, and how the kill switch revokes sessions and machines.
- Servers and the gatewayHow an MCP server sits behind Straza's gateway, which tools a session sees, and what the gateway does with each tool call.
- CredentialsWhose credential a call to an MCP server carries, and why the agent never holds it.
- Audit and evidenceWhat each audit record holds, how the hash chain shows an edit or a removal, how records reach your SIEM, and what an auditor can prove with them.