stagehand

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow ingests outsider-authored free text indirectly as the action/instruction strings passed to Stagehand methods (act(), extract(), observe()) and uses them to control/observe arbitrary webpages via a browser (e.g., via page.goto(...)), so an outsider who can supply those strings can attempt indirect prompt injection.