lightpanda-browse

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs connecting LightPanda to arbitrary websites and shows code to navigate and scrape page content (e.g., page.goto + page.evaluate in "Navigate and Scrape"), which means the agent will fetch and read untrusted third‑party web pages whose content could contain instructions that influence agent behavior.