customgpt-rag

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

• Potentially malicious external URL detected (high risk: 0.80). This skill invokes the CustomGPT MCP endpoint at runtime (the configured <YOUR_MCP_URL>, obtained via https://app.customgpt.ai) to fetch knowledge-base documents that are injected into the agent context and thus directly influence/drive prompts and responses.