CHAINGUARD WORKSPACES

Last Update: November 5, 2026

Chainguard Workspaces is a hosted development environment that allows Customers and their AI Agents to write, build, test, and run code inside isolated, Chainguard-hardened microVMs (each, a “Workspace”). Customer may sync its working trees, files, and configurations (“Secured Platform Inputs”) into a Workspace and run commands, scripts, packages, and AI Agents there. 

Customer is responsible for Secured Platform inputs and is solely responsible for all actions taken in a Workspace by its users or by any AI Agent Customer runs or connects to it whether or not Customer foresaw, reviewed, or approved them. Actions taken by an AI Agent are Customer’s actions as between the parties, whether or not Customer foresaw, reviewed, or approved them. Customer is responsible for configuring agent permissions, credentials (which should carry only minimum necessary privileges), network access, and resource limits, and for supervising agent activity. Chainguard does not control or assume responsibility for any third-party agent, model, or service Customer connects to a Workspace, or for the Secured Platform Inputs or credentials Customer makes available to it. 

Chainguard Workspaces is subject to the Secured Platform Terms: https://www.chainguard.dev/legal/secure-platform-terms as Technology Preview and is not a generally available Product or Service.