Jim Clark argues that agent safety should come from narrowly scoped capabilities rather than constant human approval. He separates the agent harness from the tools and resources it can reach, using a newsroom workflow and a coding agent's occasional need for commit-signing keys to illustrate task-specific sandboxes.
An MCP gateway provides a shared control point for exposing tools, resources and prompts across different harnesses. Jim Clark advocates keeping credentials outside agent sandboxes and progressively disclosing only the capabilities required by each task. He also describes integrating agent authorization with existing organizational identity systems. The talk presents an architectural approach rather than a demonstrated guarantee that agents cannot cause harm.
Watch on YouTube
