Chainguard Launches Repository to Secure AI-Driven Software Development

KIRKLAND, Wash. – March 17, 2026 – By Nancy Torres

In a direct response to the escalating security crisis plaguing the software supply chain, Chainguard today announced the launch of Chainguard Repository. The new platform offers a single, unified source for open source artifacts—from containers and libraries to virtual machine images—all designed to be secure by default. The launch comes as development teams grapple with the dual-edged sword of artificial intelligence, which accelerates both software creation and the sophistication of cyberattacks.

Chainguard, a company focused on hardening the open source ecosystem, is positioning its new repository as a foundational trust layer for an era where AI agents and automated tools are increasingly responsible for writing and deploying code. The platform integrates intelligent security policies directly into the consumption of open source components, aiming to shift the industry from a reactive posture of scanning and patching to a proactive model of prevention.

The AI-Fueled Supply Chain Crisis

The modern software development landscape is built on a vast foundation of open source components, but this foundation is proving increasingly unstable. The speed and scale of AI are exacerbating long-standing vulnerabilities. Industry reports paint a stark picture: in 2025 alone, over 1.2 million new malicious packages flooded major repositories like npm and PyPI, and open source downloads soared to an unprecedented 9.8 trillion. Attackers, leveraging generative AI to prototype malware and craft sophisticated phishing campaigns, have dramatically increased their operational tempo, with AI-assisted cyberattacks reportedly surging by nearly 90% in the last year.

This creates a dangerous paradox for engineering teams. While AI coding assistants promise to boost productivity, they also increase dependency consumption and can introduce subtle flaws or even recommend malicious packages. The result is a growing mountain of security debt, with recent studies indicating that 82% of organizations now harbor it, much of it originating from third-party code. With the average container image containing hundreds of known common vulnerabilities and exposures (CVEs), the choice has been to either move fast and accept immense risk, or slow innovation to a crawl for manual security reviews.

"AI is dramatically increasing the speed of software development for defenders and attackers alike," said Dan Lorenc, CEO and Co-founder of Chainguard, in a statement. "AI coding tools and autonomous agents are generating more code, pulling in more dependencies, and interacting with open source at a scale humans have never seen before. Chainguard Repository is the trust layer for this new era. In a world where software is increasingly generated and deployed autonomously, trust must be built into the foundation."

A Proactive Defense: Beyond Scanning

Chainguard Repository’s core innovation lies in its 'secure-by-default' philosophy, a stark contrast to the prevailing industry practice of using artifact managers like JFrog Artifactory or Sonatype Nexus to store components and then scan them for known vulnerabilities. While scanning is a critical security function, it is inherently reactive, identifying problems only after a potentially vulnerable component has already entered an organization’s environment.

Chainguard’s approach is preventative. Instead of just managing artifacts, the company rebuilds them from source code in a secure, isolated environment that is compliant with Supply-chain Levels for Software Artifacts (SLSA) Level 3 standards. This process ensures a verifiable and tamper-resistant chain of custody from source to binary.

The resulting artifacts are built on Wolfi, a minimal Linux "undistro" that strips out all non-essential components, including shells and package managers, to drastically reduce the attack surface. These hardened images are rebuilt daily to incorporate the latest security patches, virtually eliminating patch lag. Each artifact is also shipped with a Software Bill of Materials (SBOM) and is cryptographically signed using the Sigstore standard, allowing organizations to verify its integrity and provenance.

Starting today, the repository offers over 73,000 Chainguard-built JavaScript packages, which the company claims eliminate 99.7% of malware by design. A unique "cooldown" feature adds another layer of protection by delaying the availability of newly published upstream packages, giving the security community time to identify and flag potential attacks before they can be pulled into a corporate environment.

Enterprise Governance and the Road Ahead

For large enterprises, managing the sprawling use of open source software is a significant governance and compliance challenge. Chainguard Repository directly addresses this by embedding configurable, intelligent policies into the platform. These policies allow organizations to enforce their security standards automatically, without relying on burdensome manual reviews.

Later this year, the repository will expand its policy controls to include several key functions. CVE blocking will prevent developers from pulling artifacts with known critical vulnerabilities. License enforcement will restrict usage to a pre-approved list of licenses, mitigating legal risk. Other planned policies will block the use of end-of-life software and enforce the use of packages with long-term support, ensuring all dependencies are actively maintained.

This suite of controls provides CISOs and compliance officers with granular visibility and enforcement across their entire software stack. Dashboards will offer real-time insights into artifact coverage, policy violations, and the organization's overall vulnerability status. This aligns with a broader regulatory push, including updated NIST guidance, that calls for greater transparency and attestation in the software supply chain.

The repository is designed to integrate with existing artifact managers or function as a standalone experience. The initial beta launch focuses on JavaScript libraries, with plans to expand to Python and Java libraries, container images, OS packages, virtual machine images, and even CI/CD workflows and AI agent skills. This ambitious roadmap aims to provide a single, trusted source for the entire modern software stack, empowering both human engineers and their AI counterparts to build quickly and safely.