Chainguard launches beta tools for safer AI coding changes
Workspaces and Checks turn tests and attestations into part of the coding flow, so teams can validate changes before they reach review.
Chainguard has introduced two beta tools, Workspaces and Checks, to make AI-agent and developer code changes safer by combining isolated execution with verifiable test results. Workspaces gives each session a hardened remote microVM, syncs a local tree with uncommitted changes into it, and can preserve state when a user detaches; Checks runs CI-style steps in separate warm microVMs and can be triggered from the CLI, a pre-commit hook, GitHub App, MCP server, or SDK. Used together, they let a Workspace pause while tests run, then resume with structured failure output or a signed attestation that can be used as a merge gate instead of rerunning the suite. Chainguard says both products are in 30-day beta trials, Workspaces currently supports Claude, and pricing is based on consumed compute.
Why it matters
For teams using AI agents or local development environments, the change is that code can be checked in the same isolated setup where it was written, then backed by an attestation tied to the environment, result and commit. That gives maintainers a clearer gate for accepting changes without rerunning the same suite elsewhere, while Workspaces also starts with Claude support and charges by consumed compute.
Keep or strike?
Does this story matter, or is it hype? Mark it before you see what everyone else did.
Sources
- InfoQ