tooluniverse-variant-interpretation

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

• Third-party content exposure detected (low risk: 0.10). The workflow queries public literature databases like PubMed and EuropePMC, where search queries can retrieve outsider-authored abstracts and preprints.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill references external code repositories and URLs during setup/runtime instructions (such as https://github.com/evolutionaryscale/esm and https://forge.evolutionaryscale.ai) to install and authenticate unmerged feature-branch code and model services, qualifying as an unverifiable runtime external dependency under the dependency gate.