pq4r

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

• Third-party content exposure detected (medium risk: 0.30). Source text is ingested directly from user-supplied inputs via the /pq4r <path|URL|pasted text|transcript> runtime path and then read in Phase 3 (“User-provided/local content may be chunked directly” / third-party web material is excerpted/paraphrased), so an outsider can provide free text for the LLM to process through the required workflow.