Chainguard launches Secure Workspaces and Checks for AI coding agents

Chainguard announced two beta products: Workspaces, remote dev environments on hardened microVMs, and Checks, which moves CI into the development loop. The tools isolate code written by developers and AI agents using short-lived credentials and tamper-proof logs.

• Workspaces gives each agent a dedicated microVM with working-tree sync
• Checks runs CI steps in warm microVMs and returns structured logs
• A host-side proxy injects short-lived credentials instead of long-lived secrets
• Both products are in 30-day beta; Workspaces currently supports Claude