What is Aegisora?

Aegisora is an execution-control layer for AI agents. It sits beneath your agent orchestration layer, at the point where a request becomes an executable action, and returns an explicit decision for each action before it runs.

What it does

• Evaluates each agent action against the policies of a workspace.
• Returns one of three states: ALLOW, BLOCK or ESCALATE.
• Blocks by default when no policy matches the action.
• Returns an evidence hash with every decision so the decision can be traced downstream.

What it works with

The boundary is framework independent. It sits beneath your agent stack rather than inside one specific framework, and provider execution can pass through the same enforcement path.