unified-ai-gateway

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

• Potentially malicious external URL detected (high risk: 0.90). The skill pulls and (in the registration example) runs a remote Docker image that would execute fetched code at runtime: ghcr.io/happy520ai/unified-ai-system/mcp-server@sha256:cc17e923335f953631f59fb6a5ffcdce0e12e16c5abf362f1d28747452adadee