Chainguard introduces secure development tools for AI coding agents

Software supply chain security company Chainguard has announced two new beta products, Chainguard Workspaces and Chainguard Checks, designed to secure development environments for both human developers and AI coding agents.

Chainguard Workspaces provides remote development environments using hardened microVMs. This approach addresses security risks associated with AI agents that require permissions to install packages or run shell commands, which could otherwise expose corporate credentials on a local laptop. The system uses a host-side proxy to provide short-lived credentials for approved outbound requests and maintains tamper-proof logs of commands, code changes, and network calls.

Chainguard Checks enables continuous code verification within active coding sessions. Unlike traditional continuous integration (CI) that tests code after submission, this platform uses isolated, warm microVMs to validate software during the development loop. Successful tests generate digitally signed records, allowing approved code to move through pipelines without redundant verification. This system is intended to reduce computing expenses and accelerate development cycles while providing structured feedback to AI agents for immediate error correction.