Use case

Put a local security reviewer inside the agent loop.

MCP tools let agents ask Radar what changed, what is risky, and what should be fixed first.

Agent-safe context

The agent receives structured findings and report resources instead of guessing from raw terminal output.