claim-source-verification

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow has the “skeptic” role open and read the actual proposed source to quote supporting sentences verbatim (“No verbatim supporting sentence, no confirmation” / “quote the sentence in the source that carries this claim”), so any outsider-provided source content that is fetched/ingested as the “actual source” would be exposed for prompt injection via free text inside that document.