News
Field notes from building enforcement for coding agents, IDEs and browsers — what actually breaks, and what it costs to catch it.
9 min read
Before deploying AI coding agents across the enterprise, CISOs need controls for identity, permissions, sensitive data, production access, audit evidence, and policy enforcement. Here’s what to evaluate.
Read more →9 min read
: AI agents can generate and execute shell commands with real infrastructure consequences. Learn how enterprises can enforce policies, permissions, approvals, and audit controls before AI-generated commands execute.
Read more →6 min read
Learn the difference between AI governance and runtime enforcement, and why organizations need policy controls at the point where AI agents take action.
Read more →10 min read
AI agents are changing the enterprise security boundary. Prompt-level controls are no longer enough when AI can execute commands, modify files, invoke tools, and act against enterprise resources.
Read more →7 min read
AI governance needs more than visibility. It needs a way to make decisions when AI moves from generating an answer to taking an action.
Read more →10 min read
Prompt injection is an input problem. Agent security is an action problem.
Read more →Background on the subjects these notes assume.
