Runtime Proof
From agent action to auditable decision
Aegisora sits between your AI agents and the systems they touch. Each action is evaluated before it runs and leaves verifiable evidence behind.
Step 1
Agent action
An agent requests a tool call or data access through the Aegisora API.
Step 2
Context and risk
Aegisora enriches the request with workspace, identity and risk context.
Step 3
Policy decision
Policies return ALLOW, BLOCK or ESCALATE, with sensitive data masked.
Step 4
Evidence
Every decision is written to a hash-chained evidence record for audit.
Illustrative decision
Simplified example. See the documentation for the exact API contract.
{
"decision": "BLOCK",
"reason": "Policy: no external transfer of customer PII",
"risk": "high",
"evidence": { "recorded": true, "chained": true }
}