knowledge-discovery

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly lists and requires using public third-party datasets (e.g., WALS at wals.info and HistWords at nlp.stanford.edu) and the protocol (Data exploration, feature engineering, etc.) implies the agent will fetch and interpret that open web content, which can materially influence its analyses and actions.