Agent security startup emerges with a policy engine for autonomous software
A seed-stage company launches tooling to monitor, constrain and halt AI agents operating inside company systems.
Source: Huburb demo record · no external source attached
CybersecurityDemo data
Veritas Ops
Monitoring and containment for autonomous software agents.
What they do
Veritas records every action an internal AI agent takes, applies policy, and can halt a run that steps outside its permitted scope.
Why it's interesting
Huburb analysis — an editorial reading of the approach, not a sourced claim.
Agents are a new class of privileged actor inside company systems, and existing identity tooling was not designed for software that improvises.
Products
- Agent policy engine
- Action audit trail