team-skills

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

• Third-party content exposure detected (low risk: 0.10). The workflow downloads and installs external skills from URLs specified in a manifest file, representing low-risk retrieval of user-selected external content.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill downloads external skills over HTTPS at runtime from URLs defined in a manifest, but because the specific remote sources are arbitrary and user/manifest-defined without verified provenance, it presents an unverifiable runtime external dependency.