open-browser-use

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the LLM context via the runtime-controlled browser page content: the workflow uses open-tab/navigate/SDK tab.playwright.domSnapshot() / locator(...).inner_text() and similar actions that ingest arbitrary text from user-specified URLs or existing user tabs (public web pages or other outsiders’ content) into the agent’s context.