human-ui

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

• Third-party content exposure detected (low risk: 0.10). The workflow fetches design guidelines from a specified URL using WebFetch, which ingests external content.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to fetch external behavior-selecting guidelines from an arbitrary third-party domain (guidelines.sh) at runtime, representing an unknown/weak operational dependency.