News
Field notes from building enforcement for coding agents, IDEs and browsers — what actually breaks, and what it costs to catch it.
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 →