llm-wiki-capture

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

• Third-party content exposure detected (medium risk: 0.30). In the required Source Ingest branch, the workflow explicitly ingests “supplied external material” chosen from the user’s ingest request (including browsing/extraction of the supplied URL/content), so outsider-authored free text can be provided at runtime and consumed by the LLM without requiring the agent to pre-select a specific item beyond the user’s request intent.