Use case B · active · requires a signed mandate

Audit your AI chatbot.

Probes a chatbot or agent endpoint for system-prompt leaks, jailbreaks, token-drain abuse, tool misuse, and data exposure — then scores it. This is active testing: it only runs against a target covered by a valid authorization mandate.

Chatbot endpoint URL

Response path (JSON dot-path to reply)

Request body template — {{message}} is replaced with each probeHeaders (JSON, optional — e.g. Authorization)Your business rules (optional, one per line) — the audit tests each against your botConsistency mode — run each probe 3× to catch flaky guardrails (uses more credits/time).

Authorization — verify domain ownership to unlock the audit

1Prove you own the domain
2Accept the mandate & verify
I am authorized to order this test and I accept the audit mandate & terms. I confirm I will use a non-production, dummy-data account, and I am responsible for the accuracy of the information provided.
3Scan unlocked
Complete steps 1–2 to unlock the scan.

← Back to domain scan