recall

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md:17-25 describes mining a “shared record” across issue tracker, chat/issue channels, and error tracking, which at runtime would ingest outsider-authored free-form text (e.g., issue bodies/comments) into the agent’s LLM context via the transcript/mining and “why” source investigators.