literature-review

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

• Third-party content exposure detected (low risk: 0.10). The agent skill executes academic search queries and processes retrieved bibliographic metadata, papers, and abstracts, which are user/third-party indexed scholarly works but do not constitute unauthenticated outsider-authored submission feeds (like issues or support tickets); however, since the agent searches and fetches arbitrary web/scholarly content via APIs and search tools without a strict pre-selected item constraint, it falls under active search/fetch of external content.