browser-automation

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

• Third-party content exposure detected (low risk: 0.10). The browser automation workflow navigates to web pages and extracts or interacts with page content, exposing the agent to arbitrary outsider-authored web page text upon active fetching.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs downloading and executing a remote installation script from inference.sh (cli.inference.sh) which constitutes a weak/unknown third-party operational dependency.