summarize

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly fetches and summarizes arbitrary open-web URLs and public YouTube links (e.g., the examples summarize "https://example.com" and "https://youtu.be/..."), so it ingests untrusted, user-provided third-party content that the agent reads and interprets, enabling indirect prompt injection.