Agentic Trust Controls: ready-to-use security controls for AI agents

Agentic Trust Controls

An open-source set of 65 controls designed to extend ISO 27001, ISO 42001, and similar frameworks for agentic AI risks. It offers two baselines: Developer (43 controls) and User (22 controls), covering identity, instruction integrity, adversarial testing, and runtime monitoring across 12 domains. Backed by Vanta, the controls can be exported by role and framework; an MCP server integration is listed as coming soon. The post does not disclose the actual content of individual controls or implementation cost.