Cursor Earns AIUC-1, an AI Agent Security Certification That Covers MCP

Cursor announced on August 13, 2026 that it has earned AIUC-1 certification, a standard for AI agent security, safety and reliability that pairs an audit of organizational controls with adversarial testing of the product. Keeping it requires quarterly retesting.

Cursor AI-agent security MCP certification

Cursor announced on August 13, 2026 that it has earned AIUC-1 certification, a standard for AI agent security, safety and reliability¹. By the company’s account, AIUC-1 combines an audit of organizational controls with adversarial testing of the product itself, translating existing frameworks such as the NIST AI Risk Management Framework, MITRE ATLAS and the OWASP agentic threat taxonomy into requirements testable against live AI systems¹. For coding agents, those requirements cover secrets protection, secure code generation, MCP security, and agent identity and permissions¹.

The audit was carried out by Schellman, and Cursor says its IDE and cloud agents were put through two rounds of testing across several thousand scenarios, probing behaviors such as generating vulnerable code and running unsafe commands¹. Maintaining the certification requires testing at least quarterly plus a full audit each year, and the requirements themselves are updated quarterly¹.