playwright-mcp

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The MCP tool calls (e.g., browser_type: { selector: "#password", text: "pass" }) place plaintext secrets directly into the agent's generated commands, so using real credentials would require the LLM to emit those secret values verbatim (high exfiltration risk).

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow uses browser_navigate to load arbitrary URLs and then tools like browser_get_text / browser_snapshot / browser_evaluate read page body/text/DOM into the agent context, so outsider-authored free text from fetched web pages can be ingested via the browser-extraction path.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime relies on running the @playwright/mcp package via npx (which fetches and executes remote code from the npm package page https://www.npmjs.com/package/@playwright/mcp), so an external URL is used at runtime to fetch and run required code.