oma-security

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

• Third-party content exposure detected (medium risk: 0.30). The workflow consumes outsider-authored free text by scanning and analyzing application source code, agent skill packages, and MCP components.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill contains instructions for setting up and executing security scan engines and tools from remote GitHub repositories (such as vercel-labs/deepsec, cisco-open/ai-deep-sast, cisco-ai-defense/skill-scanner, cisco-ai-defense/mcp-scanner, and Autumn-27/ARTEX), which constitute strong official and reputable unofficial operational dependencies.