company-recording-clip

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

• Third-party content exposure detected (high risk: 0.75). The required workflow ingests outsider-authored free text from “chat or collaboration tools” (e.g., user-provided fuzzy requests/messages and any comments/threads) via Zapier MCP, which can contain arbitrary third-party text that the agent may read into the LLM context for caption search/transcription.