browser-gui-inspect

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill exposes high-risk capabilities (capture of full network request/response bodies, reuse of a shared persistent browser profile with live sign‑in sessions, dynamic npx fetch of the MCP package, and no built-in domain allowlist) that are easily abused to exfiltrate credentials or other sensitive data even if the document's intent is benign.

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 reads arbitrary, operator-supplied page content from the target URL via navigate_page followed by take_snapshot/take_screenshot (workflows 1–3), so outsider-authored free text can be ingested whenever a user points it at a page/feed that contains such text.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill invokes an external package at runtime via "npx -y chrome-devtools-mcp@latest", which fetches and executes remote code (the chrome-devtools-mcp npm package) and is required for the skill to run, so this runtime dependency is a potential execution risk.