Introducing Chainguard Harness: Proven Vulnerabilities and Ready-to-Merge Fixes From Frontier AI, No Model Access Required

Available through the Claude Marketplace by Anthropic, Harness unleashes frontier AI on a live, isolated build of an organization's application and delivers only proven vulnerabilities, each with a recommended fix ready to merge

SEATTLE, Oct. 14, 2026 /PRNewswire/ -- Chainguard, the trusted source for open source, today announced Chainguard Harness, a new cyber harness that uses frontier AI models like Anthropic's Claude Mythos 5.1 to find and fix exploitable vulnerabilities in an organization's application code. Harness builds an application from an organization's source code in an isolated microVM, instructs a frontier model to hunt for exploitable vulnerabilities in a live, running instance, and reproduces every finding before delivering a tailored, staff-reviewed fix. Teams benefit from frontier vulnerability hunting without needing their own access to frontier models, and act only on proven findings rather than sifting through noise. Chainguard Harness is now available in the Claude Marketplace, so customers can purchase Harness directly from Anthropic and consolidate their AI spend. 

Finding vulnerabilities is no longer the hard part

Frontier AI models have become remarkably effective at finding and chaining exploitable vulnerabilities, but most organizations cannot put that capability to work on their own code. Access to frontier models that perform vulnerability discovery across first-party and open source code has been mostly available to a select group of organizations performing defensive security work for critical infrastructure. Traditional scanners add to the problem: they analyze static code, flag issues that may never be reachable, and miss the logic flaws, injection paths, and resource exhaustion bugs that only appear when an application is actually running. The result is a growing backlog of unverified alerts, with engineers spending hours and tokens triaging findings and writing their own patches instead of shipping. Chainguard Harness closes the gap by finding exploitable, reproducible vulnerabilities and providing fixes for both open source and first-party code.

"Every time we've pointed Harness at code, it has found serious, exploitable vulnerabilities, including in software that's been scanned many times over and code that runs critical infrastructure," said Dan Lorenc, CEO and Co-founder, Chainguard. "That capability is beginning to make its way to more and more organizations, and the teams that move first will be the ones that find and fix these issues before they become incidents. Harness offers another option for security teams to get proven vulnerabilities and fixes their engineers can ship."

Proven vulnerabilities, with fixes written for your code

Chainguard Harness connects to a team's repository through the Guardener GitHub app, builds the application exactly as it ships, and boots it in an isolated microVM. A frontier model then probes the live application for exploitable vulnerabilities, testing it in context, the way an attacker would, rather than reading static code. Every finding is replayed in a fresh, isolated environment until it reproduces every time, so teams only see what is proven. When a vulnerability is confirmed, Chainguard delivers reliable reproduction steps alongside a fix written for that specific codebase.

Chainguard Harness includes:

• Frontier AI without frontier access: Teams get the benefit of frontier-model vulnerability hunting on their own applications without needing their own access to frontier models.
• Built for the bugs that live at runtime: Harness tests a running copy of the application instead of just static code, surfacing logic flaws, injection, and resource exhaustion that only show up in context.
• Isolation by design: Every run happens in its own microVM sandbox built from source on Chainguard OS, with no network egress, no credentials inside, and full audit logging.
• Reproducible, verified findings: Every finding is replayed in a fresh, isolated environment until it lands every time, reducing false positives and security noise.
• Fixes written for your code: Custom recommendations across first-party code and open source dependencies give teams a path to secure code, not a list of risks.
• Human review on every fix: A Chainguard engineer reviews and approves every fix and report before it reaches a customer.
• Responsible disclosure built in: Vulnerabilities found in open source go upstream to maintainers through Athena, Chainguard's coordinated vulnerability disclosure process.

At launch, Chainguard engineers operate Harness on behalf of customers by running it against the trusted open source artifacts Chainguard already distributes and delivers patched builds via Chainguard Repository, and by running it against a customer's source code and delivering an engineer-reviewed report of validated findings with severity, impact, reproduction steps, and recommended remediation.

Available through the Claude Marketplace by Anthropic

Chainguard Harness is available through the Claude Marketplace by Anthropic, allowing enterprises to use a portion of their existing Anthropic commitment to procure it. Chainguard Harness reproduces vulnerabilities discovered by Mythos 5.1 in an isolated environment, ensuring every fix is reviewed by a Chainguard engineer, and sharing results that customers can act on immediately. Anthropic manages invoicing, so Anthropic customers can add Harness without a separate procurement process and consolidate their AI spend.

"Models like Claude Mythos 5.1 can surface vulnerabilities that have gone unnoticed for years, so how that capability is deployed matters as much as the capability itself," said Daniela Hurtado, Head of Commercial Cyber, Anthropic. "Chainguard Harness surrounds Mythos 5.1 with isolation, reproduction, and human review, so defenders get the benefit, and every finding is verified before it reaches a customer. Through the Claude Marketplace, enterprises can put their existing Anthropic commitment directly toward that kind of risk reduction."

Chainguard Harness is available through a beta waitlist. To learn more, visit www.chainguard.dev/harness.

About Chainguard
Chainguard is the trusted source for open source. By providing engineers and AI agents with hardened, trusted, and production-ready artifacts, Chainguard helps organizations prevent AI supply chain attacks, increase engineering velocity while reducing toil, and maintain continuous compliance. Customers include Fortune 500 enterprises and global industry leaders, including Anduril, Canva, DocuSign, OpenAI, Public Storage, Snap Inc., and Snowflake. Chainguard is venture-backed by leading investors, including Amplify, IVP, Kleiner Perkins, Lightspeed Venture Partners, Mantis VC, Redpoint Ventures, Sequoia Capital, and Spark Capital.  For more information, visit: https://www.chainguard.dev/ 

Media Contact: Brittany Hendrickson, [email protected] 

SOURCE Chainguard