cloud

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). Documentation for a browser-automation cloud service that includes instructions and capabilities for stealth browsing, residential proxies, CAPTCHA/Cloudflare bypass, embedding API keys in connection URLs, profile-sync scripts, and sharing live browser sessions — which can facilitate evasion and automation misuse but is not explicitly malicious.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (medium risk: 0.30). The docs instruct running an unverified remote install script via "curl -fsSL https://browser-use.com/profile.sh | sh", which downloads and executes code directly from browser-use.com.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s required runtime workflow ingests outsider-authored free text when a caller supplies the task prompt (and potentially systemPromptExtension/agent_prompt for skills) into Cloud API v2/v3 endpoints like POST /tasks or client.run() at runtime.