Project Glasswing: Securing critical software for the AI era

[Hamartia Antidote]

Project Glasswing: Securing critical software for the AI era

A new initiative to secure the world’s most critical software and give defenders a durable advantage in the coming AI-driven era of cybersecurity.

www.anthropic.com

Today we’re announcing Project Glasswing​

We formed Project Glasswing because of capabilities we’ve observed in a new frontier model trained by Anthropic that we believe could reshape cybersecurity. Claude Mythos2 Preview is a general-purpose, unreleased frontier model that reveals a stark fact: AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities.

Mythos Preview has already found thousands of high-severity vulnerabilities, including some in every major operating system and web browser. Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely. The fallout—for economies, public safety, and national security—could be severe. Project Glasswing is an urgent attempt to put these capabilities to work for defensive purposes.

As part of Project Glasswing, the launch partners listed above will use Mythos Preview as part of their defensive security work; Anthropic will share what we learn so the whole industry can benefit. We have also extended access to a group of over 40 additional organizations that build or maintain critical software infrastructure so they can use the model to scan and secure both first-party and open-source systems. Anthropic is committing up to $100M in usage credits for Mythos Preview across these efforts, as well as $4M in direct donations to open-source security organizations.

Project Glasswing is a starting point. No one organization can solve these cybersecurity problems alone: frontier AI developers, other software companies, security researchers, open-source maintainers, and governments across the world all have essential roles to play. The work of defending the world’s cyber infrastructure might take years; frontier AI capabilities are likely to advance substantially over just the next few months. For cyber defenders to come out ahead, we need to act now.

Cybersecurity in the age of AI​

The software that all of us rely on every day—responsible for running banking systems, storing medical records, linking up logistics networks, keeping power grids functioning, and much more—has always contained bugs. Many are minor, but some are serious security flaws that, if discovered, could allow cyberattackers to hijack systems, disrupt operations, or steal data.

We have already seen the serious consequences of cyberattacks for important corporate networks, healthcare systems, energy infrastructure, transport hubs, and the information security of government agencies across the world. On the global stage, state-sponsored attacks from actors like China, Iran, North Korea, and Russia have threatened to compromise the infrastructure that underpins both civilian life and military readiness. Even smaller-scale attacks, such as those where individual hospitals or schools are targeted, can still inflict substantial economic damage, expose sensitive data, and even put lives at risk. The current global financial costs of cybercrime are challenging to estimate, but might be around $500B every year.

Many flaws in software go unnoticed for years because finding and exploiting them has required expertise held by only a few skilled security experts. With the latest frontier AI models, the cost, effort, and level of expertise required to find and exploit software vulnerabilities have all dropped dramatically. Over the past year, AI models have become increasingly effective at reading and reasoning about code—in particular, they show a striking ability to spot vulnerabilities and work out ways to exploit them. Claude Mythos Preview demonstrates a leap in these cyber skills—the vulnerabilities it has spotted have in some cases survived decades of human review and millions of automated security tests, and the exploits it develops are increasingly sophisticated.

Ten years after the first DARPA Cyber Grand Challenge, frontier AI models are now becoming competitive with the best humans at finding and exploiting vulnerabilities. Without the necessary safeguards, these powerful cyber capabilities could be used to exploit the many existing flaws in the world’s most important software. This could make cyberattacks of all kinds much more frequent and destructive, and empower adversaries of the United States and its allies. Addressing these issues is therefore an important security priority for democratic states.

Although the risks from AI-augmented cyberattacks are serious, there is reason for optimism: the same capabilities that make AI models dangerous in the wrong hands make them invaluable for finding and fixing flaws in important software—and for producing new software with far fewer security bugs. Project Glasswing is an important step toward giving defenders a durable advantage in the coming AI-driven era of cybersecurity.

Identifying vulnerabilities and exploits with Claude Mythos Preview​

Over the past few weeks, we have used Claude Mythos Preview to identify thousands of zero-day vulnerabilities (that is, flaws that were previously unknown to the software’s developers), many of them critical, in every major operating system and every major web browser, along with a range of other important pieces of software.

In a post on our Frontier Red Team blog, we provide technical details for a subset of these vulnerabilities that have already been patched and, in some cases, the ways that Mythos Preview found to exploit them. It was able to identify nearly all of these vulnerabilities—and develop many related exploits—entirely autonomously, without any human steering. The following are three examples:

• Mythos Preview found a 27-year-old vulnerability in OpenBSD—which has a reputation as one of the most security-hardened operating systems in the world and is used to run firewalls and other critical infrastructure. The vulnerability allowed an attacker to remotely crash any machine running the operating system just by connecting to it;
• It also discovered a 16-year-old vulnerability in FFmpeg—which is used by innumerable pieces of software to encode and decode video—in a line of code that automated testing tools had hit five million times without ever catching the problem;
• The model autonomously found and chained together several vulnerabilities in the Linux kernel—the software that runs most of the world’s servers—to allow an attacker to escalate from ordinary user access to complete control of the machine.

We have reported the above vulnerabilities to the maintainers of the relevant software, and they have all now been patched. For many other vulnerabilities, we are providing a cryptographic hash of the details today (see the Red Team blog), and we will reveal the specifics after a fix is in place.

Evaluation benchmarks such as CyberGym reinforce the substantial difference between Mythos Preview and our next-best model, Claude Opus 4.6:

 

[Hamartia Antidote]

Anthropic scales Claude Mythos to critical infrastructure in 15+ countries | TechCrunch

Anthropic is expanding Project Glasswing, its security vulnerability program, and access to Mythos to 150 organizations across 15 countries — targeting critical infrastructure in power, water, healthcare, and communications where a cyberattack could affect 100 million people.

techcrunch.com

Anthropic scales Claude Mythos to critical infrastructure in 15+ countries​

Anthropic is expanding Project Glasswing, its joint industry initiative to find and fix critical software vulnerabilities using AI, to about 150 new organizations across more than 15 countries, the company said Tuesday.

The news comes a day after Anthropic said it had filed confidentially for an initial public offering, following a $65 billion funding round at a nearly $1 trillion valuation.

Anthropic’s Claude Mythos is at the heart of Project Glasswing. The AI firm dubbed the model its most powerful yet, able to identify thousands of zero-day vulnerabilities over several weeks. In early April, Anthropic gave 50 initial partners, including the U.S. government, access to Claude Mythos Preview to scan their codebases for vulnerabilities and security flaws.

The expanded list of organizations with access to Mythos as of today covers power, water, healthcare, communications, and hardware — industries that weren’t “well-represented” in Anthropic’s initial cohort, the company said. Many who will now have access are companies or nonprofits that maintain codebases which other organizations and governments rely upon, Anthropic noted in the blog post.

“What each partner has in common is that a successful attack on their codebase could be catastrophic,” the company said. “For most partners, we estimate that a major attack could affect more than 100 million people, with important ramifications for both global and national security.”

The expanded group includes organizations in countries friendly to the U.S., including Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea, according to Financial Times, citing a person familiar with the matter.

The FT also reported several organizations that have been given access to Mythos, including: U.S.-based identity and security management tool Okta; South Korean companies Samsung, SK Hynix, and SK Telecom; NATO, the U.S.-led military alliance headquartered in Brussels; and the EU’s cybersecurity agency ENISA.

TechCrunch has reached out to Anthropic to confirm.

Anthropic has said it expects other AI companies to soon develop models as capable as Mythos Preview, which is why the firm is racing to establish safeguards within Project Glasswing.

Since releasing Mythos, rival OpenAI released its own cybersecurity-focused model GPT-5.5-Cyber, which it has rolled out to a large group of partners for testing.

 

[Hamartia Antidote]

Project Glasswing: what Mythos showed us

In recent weeks, we pointed Mythos and other security-focused LLMs at live code across critical parts of our infrastructure. We share what we observed, the models’ strengths and weaknesses, and what the work around them needs to look like before any of it can scale.

blog.cloudflare.com

Project Glasswing: what Mythos showed us​

For the last few months, we've been testing a range of security-focused LLMs on our own infrastructure. These LLMs help identify potential vulnerabilities in our own systems, so we can fix them – and they also show us what attackers are going to be able to do with the latest models.

None of these LLMs has captured more attention than Mythos Preview, from Anthropic. A few weeks ago, we were invited to use Mythos Preview as part of Project Glasswing. We soon pointed it at more than fifty of our own repositories – to see what it would find, and to see how it works.

This post shares what we observed, what the models did well and what they didn't, and how the architecture and process around them needs to change, so they can be used at scale.

What changed with Mythos Preview​

Mythos Preview is a real step forward, and it's worth saying that plainly before getting into anything else. We've been running models against our code for a while now, and the jump from what was possible with previous general-purpose frontier models to what Mythos Preview does today is not just a refinement of what came before.

It's a different kind of tool doing a different kind of work, and that makes a clean apples-to-apples comparison to earlier models difficult. So rather than trying to benchmark Mythos Preview against general-purpose frontier models, it's more useful to describe what it can actually do, and two features that stood out across the work we did with Mythos Preview:

• Exploit chain construction - A real attack rarely uses one bug. It chains several small attack primitives together into a working exploit. For instance, it might turn a use-after-free bug into an arbitrary read and write primitive, hijack the control flow, and use return-oriented programming (ROP) chains to take full control over a system. Mythos Preview can take several of these primitives and reason about how to combine them into a working proof. The reasoning it shows along the way looks like the work of a senior researcher rather than the output of an automated scanner.
• Proof generation - Finding a bug and proving it's exploitable are two different things, and Mythos Preview can do both. It writes code that would trigger the suspected bug, compiles that code in a scratch environment, and runs it. If the program does what the model expected, that's the proof. If it doesn't, the model reads the failure, adjusts its hypothesis, and tries again. The loop matters as much as the bugs it finds, because a suspected flaw without a working proof is speculation, and Mythos Preview closes that gap on its own.

Some of what we describe above is not entirely unique to Mythos Preview. When we ran other frontier models through the same harness, they found a fair number of the same underlying bugs, and in some cases they got further than we expected on the reasoning side too. Where they fell short was at the point of stitching the pieces together. A model would identify an interesting bug, write a thoughtful description of why it mattered, and then stop, leaving the actual chain unfinished and the question of exploitability open. What changed with Mythos Preview is that a model can now take those low-severity bugs (which would traditionally sit invisible in a backlog) and chain them into a single, more severe exploit.

Model refusals in legitimate vulnerability research​

The Mythos Preview model provided by Anthropic, as part of Project Glasswing, did not have the additional safeguards that are present in generally available models (like Opus 4.7 or GPT-5.5).

Despite this, the model organically pushes back on certain requests - much like the cyber capabilities that made it useful for vulnerability hunting, the model has its own emergent guardrails that sometimes cause it to push back on legitimate security research requests. But as we found, these organic refusals aren’t consistent - the same task, framed differently or presented in a different context, could produce completely different outcomes as illustrated in the examples below.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

First findings from Project Glasswing