Chainguard puts Claude Mythos on flaw hunting, secures coding agents and widens Libraries

Chainguard puts Claude Mythos on flaw hunting, secures coding agents and widens Libraries

Secure software supply chain solution provider Chainguard Inc. today unveiled a vulnerability-hunting service that uses frontier artificial intelligence models such as Anthropic PBC’s Claude Mythos 5.1, as well as two products for securing AI coding agents and an expansion of Chainguard Libraries.

All four arrived at the company’s Assemble London conference today.

Vulnerability hunting falls to Chainguard Harness, which reports only the flaws it can reproduce against a live build of a customer’s application. Developers and their agents get a hardened, credential-free cloud environment from Chainguard Workspaces.

Because that work still has to be tested, Chainguard paired Workspaces with Chainguard Checks, which runs continuous integration tests on the code before anyone opens a pull request. Go and .NET developers are the new audience for Chainguard Libraries, the company’s service for vetted open-source dependencies.

Harness is aimed at organizations that want frontier-model vulnerability hunting without direct access to the models themselves. According to Chainguard, those models have mostly gone to a small set of groups doing defensive security work, and static scanners miss the logic flaws that only surface once an application is running.

Customers connect a repository through Chainguard’s Guardener GitHub app. From there, Harness compiles the application as it would ship, and the build boots in an isolated virtual machine cut off from the outside network. A frontier model then probes that running copy the way an attacker would. Nothing gets reported until it reproduces in a fresh environment on every replay, according to the company. A Chainguard engineer signs off on every report and fix before a customer sees it.

“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 co-founder and Chief Executive Dan Lorenc.

For now, Chainguard’s own staff run Harness for customers. Scans of the open-source artifacts the company distributes end in patched builds shipped through Chainguard Repository.

When Harness runs against a customer’s own code, the customer receives a write-up covering severity, reproduction steps and a recommended fix. Flaws turned up in open-source projects go to upstream maintainers through Athena, a disclosure coalition Chainguard set up in June.

Anthropic is also selling Harness. The service is listed on the Claude Marketplace, which opened in March, and enterprises can put part of an existing Anthropic spending commitment toward it. Daniela Hurtado, head of commercial cyber at Anthropic, said how a model like Mythos 5.1 gets deployed “matters as much as the capability itself.”

Harness looks for flaws in software a company already ships. The two agent products step in earlier, at the point where coding agents are still writing the code. Chainguard Workspaces and Chainguard Checks are built for teams whose agents run long migrations with the same permissions as the developer who launched them, within reach of cloud keys and production systems. Approval prompts depend on the agent following the rules, and a developer juggling a dozen agents can end up clicking “allow” all day.

Each workspace is a remote virtual machine built entirely from Chainguard OS packages. Syncing with the developer’s local working tree keeps the editor and git workflow on the laptop while execution happens remotely.

Real credentials stay outside the machine. Tools that need one on disk get an encrypted, short-lived stand-in, and teams can switch on egress policies that block any connection they have not approved. Lorenc described the result as an isolated machine “with nothing in it worth stealing.”

Checks takes aim at the queue that forms when agents push code far more often than continuous integration systems were designed for. Each run starts from a warm snapshot with dependencies already installed. An agent working in a workspace has its session suspended during testing, then picks up again with any failure logs fed straight into its context.

Passing runs produce a signed attestation tied to the environment, test result and commit. Merge gates can verify that attestation and merge the code without running the suite again.

James Governor, co-founder of analyst firm RedMonk LLC, said existing continuous integration platforms are “struggling under the massive increase in load” that agentic development has brought. Australian payments infrastructure provider Cuscal Ltd. is an early design partner on both products.

The Libraries update takes the service to five languages. Java was the only language Libraries supported when Chainguard introduced the service in March 2025, and Python and JavaScript have joined it since. Teams building on .NET can now swap NuGet for a Chainguard-managed endpoint.

Go and .NET both come with configurable blocking policies and malware scanning that also catches greyware. Chainguard applies the term to packages that work as advertised while also doing something malicious, such as harvesting credentials.

Chainguard is pitching the Go libraries at teams that build cloud and infrastructure software in the language. Older Go packages also get backported upstream fixes when a critical- or high-severity known vulnerability is found, sparing teams pinned to them a forced major upgrade. JavaScript libraries now carry the same remediation.

Java and Python get zero-day protection drawn from findings that Athena members pool. Chainguard says patched artifacts for flaws found by frontier AI models will be ready the day each one is made public.

Workspaces and Checks are available in beta test starting today. Harness is open through a waitlist for its own beta, and early access has started for the Go and .NET libraries.

Chainguard is a venture capital-backed startup that has raised about $892 million since its founding in 2021. The latest $280 million came from General Catalyst’s Customer Value Fund last October.

Image: SiliconANGLE/GPT Image 2.5