Chainguard, a software supply chain security company, is taking its Agent Skills platform to the next level. This platform is designed to secure the fast-growing world of AI coding agents by providing a public registry of over 1,000 hardened agent skills, a private registry, and a hardening service for internal, organization-specific skills. Chainguard's Agent Skills platform is a continuously maintained catalog of hardened AI agent skills, aiming to bring "secure by default" practices to the emerging agent ecosystem. The new update turns Agent Skills into both a public clearinghouse of secured community skills and a home for an organization's internal skills. It also provides a hardening-as-a-service tier for teams that want Chainguard to do the heavy lifting to ensure their homebrew agents are safe. Chainguard's hardening pipeline scans public skills against a ruleset designed to catch common and emerging attack patterns, including over-permissioned scopes and capabilities, obfuscated commands and base64 execution, credential harvesting behavior, and downloads from untrusted or suspicious domains. The hardening process is treated as a continuous process rather than a one-time static approval gate. When the ruleset detects a problem, the system uses AI to rewrite and harden the skill, and each hardened skill ships with a HARDENING.md document that serves as an audit log. Chainguard is also addressing the growing sprawl of internal agent skills within organizations by providing internal skills with a proper registry namespace. Skills live at skills.cgr.dev/
