Coalition For Secure AI

The Coalition for Secure AI is an industry collaborative initiative focused on establishing best practices and technical standards for AI system security and resilience. The group brings together technology companies, researchers, and government stakeholders to address vulnerabilities in AI development and deployment that could pose risks to organizations and society. This coalition matters for enterprise governance because it provides frameworks and recommendations that help companies align their AI security practices with evolving industry consensus and emerging threat landscapes.

1 item

ResearchGlobal2026-08-18

CoSAI Paper Sets Token-Exchange Standard for Agentic Trust Boundaries

The Coalition for Secure AI has published guidance establishing token exchange at every agent trust boundary as a foundational control principle for agentic workflows. The paper addresses credential risk in agent-to-tool and agent-to-agent handoffs, where standing credentials create outsized exposure. It offers compliance teams a concrete authorization model to apply to autonomous AI deployments.

Coalition for Secure AI agentic AI token exchange identity and access management agent authorization trust boundaries credential governance