Tool count is a vanity metric. Annotation coverage is what makes an AI agent safe.

The focus on tool count when evaluating AI agents is misleading; what truly determines safety is annotation coverage. Annotation coverage provides crucial information about each tool's impact and behavior, ensuring the agent doesn't make unsafe decisions. This insight highlights the importance of understanding how AI agents interact with tools rather than just counting the tools available. It's a reminder that meaningful oversight and safety come from detailed knowledge, not superficial metrics.