product-manager-toolkit

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

• Third-party content exposure detected (medium risk: 0.65). Outsider-authored free text (e.g., customer interview transcripts) can be read at runtime via scripts/customer_interview_analyzer.py by opening the user-supplied <interview_file.txt> and passing its contents (interview_text) into analysis/formatting that becomes LLM-readable output; then SKILL.md suggests using that resulting text in an agent prompt (which would allow prompt-injection if the transcript contains malicious text).