browser

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

• Third-party content exposure detected (high risk: 0.85). High: the required workflow navigates to a user-supplied URL at runtime (scripts/nav.js → Page.navigate), then runs Runtime.evaluate (scripts/eval.js / scripts/pick.js) and returns extracted DOM text/attributes, which can include outsider-authored free text from the loaded page into the agent’s LLM context.