browser-testing-with-devtools

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to use Chrome DevTools MCP to read live DOM content, console logs, and network responses from web pages (browser content) — untrusted third-party data the agent ingests and that can materially influence decisions (see "Available Tools" and "Security Boundaries").

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's .mcp.json instructs runtime execution of a remote npm package via npx ("@anthropic/chrome-devtools-mcp@latest"), which will fetch and execute remote code and is required for the skill to function.