triage-nda

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). The required workflow ingests the NDA text via user-provided inputs (file upload/URL/pasted text) and then analyzes it; if that NDA text originates from an outsider (e.g., the counterparty’s NDA), it becomes readable prose in the agent’s LLM context, creating indirect prompt-injection exposure.