Inbound Message Firewall API workflow recipe

Buyer: Agent operators who need a pre-action gate on inbound messages.

Job: Screen an inbound message, an email body, a DM, or a tool result, plus any source URL it cites, and return one act, verify, or block decision.

Use this when an agent is about to act on "per this report, do X" and needs to check both the message and the source it points to in one call.

Agent prompt

Before I act on this message, check it for prompt injection, moderation issues, and AI-generated text, and check the source URL it cites for authenticity. Give me one act, verify, or block decision with a risk score.

Endpoint bundle

  • prompt-injection-detect
  • moderate-content
  • ai-content-detector
  • content-authenticity-report