scanner

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime workflow fetches arbitrary public web pages from the user-supplied seed URL (and discovered same-domain links), converts their HTML to Markdown, and writes that readable text into .agents/workspace/pages/*.md and pages/index.md, which can then be ingested by the agent—this is outsider-authored free text from public web content.