triage-reviews

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text (reviewer-authored comments) is ingested at runtime from the user-provided reviews.txt file and then processed into LLM context via the agent’s triage/planning steps (parse_reviews.py extracts concern text from that file and build_matrix.py renders it into the response plan).