Governing AI agents on the endpoint: a CISO's field guide

Your developers run AI agents, and most security programmes have no control that sees them. This 12-page guide covers what an agent on a laptop can reach, why EDR, the proxy and DLP miss it, the four decisions to make, seven policy clauses to add, the managed settings for Claude Code, Codex CLI and Cursor, the evidence an auditor accepts, a framework map and a 30, 60 and 90 day plan. Leave a work email and it is sent to you.