chrome-devtools

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt shows examples that pass plaintext secrets directly on the command line (e.g., --value "secret" for passwords) and instructs constructing/issuing those commands, which requires the agent to handle and potentially emit secret values verbatim (exfiltration risk).

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow can fetch and ingest arbitrary public web page content via page.goto(args.url) (e.g., navigate.js, screenshot.js, snapshot.js, console.js, network.js, performance.js), and then reads outsider-authored free text from the page into LLM context through JSON fields like title, elements[].text, messages[].text, and result.