ce-riffrec-feedback-analysis

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

• Third-party content exposure detected (high risk: 0.75). The analyzer ingests the operating user-provided Riffrec zip/video/audio/notes, extracts readable transcript text (and meeting-notes body) from that media/markdown, and writes it into analysis.md/review-prompt.md—which are then loaded by downstream agents; if the recording/notes contain outsider-authored free text, that text becomes LLM context via the transcript/notes fields.