Every agent request — who called what, with which credential, and whether it was allowed.

The request log lists raw HTTP traffic. This reframes it as an agent egress audit trail: each row is attributed to a named agent, scored for sensitive data, and resolved to a verdict you can act on.

search agent or destination

12,480 requests today·23 flagged·14 blocked

Agent Requests

Click a row to inspect.

TimeAgentDestinationMethod·StatusTokenSensitiveVerdict

Request Detail

support-copilot

14:02:11 today

FLAGGED

Full URL

POST https://api.unapproved-gemini.example/v1/generateContent

Request

4.2 KB JSON · customer support transcript in prompt body

Response

200 · 6.8 KB completion

Credential

TypeAPI key (Bearer)

Tokensk-live-····G7

Sourcesupport-copilot env · unmanaged

Sensitive Data (qscan)

detected

12 PII matches: 8× email, 3× full name, 1× phone number found in request body.

Policy That Matched

no policy matched — destination is unapproved

Create policy from this request →