AIUC-1

Cursor is now AIUC-1 certified.

Cursor joins companies including Harvey, ElevenLabs and UiPath in earning AIUC-1, the trust mark for AI agents.

Get in touch to understand AIUC-1.

I am interested in...

The standard built for AI agents

AIUC-1 is the industry-driven agent standard designed to unlock enterprise adoption of AI. It is built with 250+ Fortune 1000 CISOs and risk leaders, updating its 50+ requirements every quarter as agents evolve.

“AIUC-1 included several thousand test scenarios, helping us validate that we’re mitigating real adversarial risk.”
Kenneth Thomas MorasSecurity, Cursor

Certified category leaders

Cursor
Lovable
Harvey
ElevenLabs
UiPath
Fin AI

Covering risks specific to coding agents

Coding is the dominant enterprise AI use case, six times larger than the next category. This quarter, AIUC-1 expanded to cover it, updating 8 requirements and 41 controls.

Secrets management

A new mandatory requirement (A008) to detect and prevent secret leakage across AI system inputs, outputs, logs and credential storage.

Secure defaults in code

A new mandatory requirement (B010) promoting secure patterns in generated code, from common vulnerability classes to safe dependency specification.

Runtime containment

Execution-level safeguards extended to coding agents, covering sandboxed execution and scanning agent configuration for prompt injection (B006.3).

More in the Q3-2026 standard update.

Built with the AIUC-1 Consortium

The coding expansion was shaped by more than 100 Consortium members over a month-long process, including CISO roundtables, technical sessions and peer review.

See our full list of Consortium members

Get in touch to understand AIUC-1

I am interested in...