mineru

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly provides a crawl command that "accepts any HTTP/HTTPS URL" and web-page extraction (and extract can take remote URLs), so the agent will fetch and parse arbitrary public web pages (untrusted third-party content) which could contain instructions that influence subsequent actions.