dogfood

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly requires a user-provided "Target URL" and — in Phase 2 of SKILL.md — instructs the agent to run browser_navigate/browser_snapshot/browser_vision/browser_console on that URL (e.g., browser_navigate(url="https://example.com/page")), so it fetches and interprets arbitrary public web pages whose content can influence subsequent clicks, typing, and reporting actions.