issue-investigation

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

• Third-party content exposure detected (high risk: 0.75). This skill ingests outsider-authored issue text (including “assignees / comments” and the reported description) and also directs runtime fetching/reading of potentially outsider-provided URLs (“Fetch the URL on the live site”, “Read the live page”), which would feed outsider free text into the LLM context during reproduction.