browser-use

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill exposes multiple high-risk capabilities that can be abused for deliberate data exfiltration and remote access — notably cookie/profile export & cloud sync (session tokens), access to the user's real Chrome profile, remote/cloud browser control and agent tasks, a background server control surface, and arbitrary Python execution within sessions — which together enable credential theft and backdoor-like abuse if misused.

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via browser-use run / browser-use extract, which uses an LLM to process page content from runtime browser navigation (e.g., arbitrary URLs opened with browser-use open <url>), including any outsider-authored text on those pages.