kimi-webbridge

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

• Third-party content exposure detected (low risk: 0.10). The skill operates a browser automation tool that navigates arbitrary web pages and reads page content (via snapshot/evaluate), exposing the agent to outsider-authored text on any website it visits.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill references a remote archive URL (https://kimi-web-img.moonshot.cn/webbridge/latest/extension/kimi-webbridge-extension.zip) used during setup/troubleshooting to obtain extension code that the user is instructed to load into the browser, representing a common/reputable unofficial operational dependency.