prove

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

• Third-party content exposure detected (low risk: 0.10). The workflow uses WebSearch, WebFetch, and external search tools during the research phase, allowing ingestion of outsider-authored web content.

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 (elan-init.sh) from an arbitrary GitHub repository (raw.githubusercontent.com/leanprover/elan/master/elan-init.sh), which constitutes an external runtime dependency with weak/unknown provenance.