browser-debugging

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes explicit plaintext credential examples (e.g., "SecurePass123!" used in mcp__claude-in-chrome__form_input) which instruct the agent to emit secret values verbatim in generated commands/actions, creating an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow uses Claude-in-Chrome to navigate to a (potentially user-supplied) URL and then ingests readable page text/DOM and console/network messages into the agent’s LLM context, which can include outsider-authored free text from public web pages or other non-user-controlled content.