Docker releases AI Governance trust layer for agents

docker.com8 min read

Docker AI Governance introduces a trust layer for agentic AI through Sandbox Policies, MCP Tool Governance, and Audit + Visibility to secure agent execution environments. The approach isolates agents within lightweight microVMs, enforces cedar-based policies on Model Context Protocol tools, and records structured audit logs. Additionally, Docker Hardened Images provide signed supply-chain provenance for container artifacts used in agent workflows. Open article →