group-search
Warn
Audited by Snyk on May 12, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (high risk: 0.90). The skill uses the yuque-mcp tool yuque_get_doc (calls to the Yuque server / yuque-mcp endpoints) at runtime to fetch external document content and inject/quote that content into the model's response, so the fetched remote content can directly control prompts/output.
Issues (1)
W012
MEDIUMUnverifiable external dependency detected (runtime URL that controls agent).
Audit Metadata