mcp-builder

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

• Third-party content exposure detected (low risk: 0.10). The evaluation harness (scripts/evaluation.py) uses WebFetch or API calls to load external SDK READMEs from GitHub, but this constitutes low-risk active fetching of selected URLs rather than unrequested monitoring of external outsider-authored feeds.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to fetch remote SDK documentation from GitHub via WebFetch (https://raw.githubusercontent.com/modelcontextprotocol/...), constituting a weak/unknown external runtime dependency.