monte-carlo-prevent

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes the Monte Carlo MCP server (https://mcp.getmontecarlo.com/mcp) at runtime to fetch table metadata, alerts, lineage and monitors which are then directly used to shape agent prompts, syntheses, and code-generation decisions, and the skill depends on that external service for its required workflows.