Chainguard launches AI flaw hunting and agent security tools

Chainguard unveiled four products at its Assemble London conference. Chainguard Harness uses frontier AI models including Anthropic's Claude Mythos 5.1 to hunt vulnerabilities in live application builds, reporting only flaws it can reproduce in an isolated virtual machine. A Chainguard engineer reviews every finding before customers see it. Chainguard Workspaces provides hardened, credential-free cloud environments for coding agents, while Chainguard Checks runs continuous integration tests before pull requests open. Chainguard Libraries expanded to support Go and .NET developers, bringing its total to five supported languages. Harness is also listed on Anthropic's Claude Marketplace.