research-interviewer

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

• Third-party content exposure detected (high risk: 0.70). The required runtime workflow includes “For analysis, invoke the interview-synthesizer subagent via Task tool” with a prompt that “Include the interview transcript and research question”; the transcript is user-provided free-form interview content (potentially authored by an outsider interviewee), which the LLM ingests as readable text into the synthesizer context.