uhi-consultation

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow involves monitoring or consuming feeds and callbacks (such as search results, callbacks, and chat messages) from external HSPA/EUA participants on the UHI network where outsider-authored free text can be submitted.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill references GitHub repositories for utility cloning and signing code without an enforced digest or immutable identifier, representing weak unofficial operational dependencies.