Project Glasswing

Anthropic is granting the E.U.'s cybersecurity agency, ENISA, access to its restricted AI security model. ENISA will participate in Project Glasswing, a program to test the vulnerability-finding model ahead of its broader release.

Anthropic announced the restoration of access to its Fable 5 and Mythos 5 AI models after the Department of Commerce lifted export controls. The Trump administration had previously imposed restrictions requiring Anthropic to suspend access for foreign users, leading to a dispute that prompted company officials to negotiate in Washington D.C.

An increase in cyberattacks on open-source software highlights challenges in securing publicly available code, with experts citing underinvestment and maintenance issues. Governments under different administrations have had mixed impacts, while companies also face criticism for insufficient responsibility. Project Glasswing identified thousands of vulnerabilities in open-source projects, but only a small fraction have been patched.

Anthropic's Mythos AI model identified vulnerabilities in classified U.S. government systems during a test with intelligence agencies. The testing, part of Anthropic's Project Glasswing initiative, raised concerns about national security risks, while tensions grew between the company and the Trump administration over AI usage restrictions.

Anthropic's Mythos model identified vulnerabilities in classified US government systems during a testing exercise with intelligence agencies via Project Glasswing. Senator Mark Warner cited the model's rapid detection capabilities, while tensions with the Trump administration led to restrictions on its AI models.

Intelligence agencies from the Five Eyes alliance (US, Canada, UK, Australia, New Zealand) warn that advanced AI models capable of significantly impacting cybersecurity are months away from public availability. These models, including Anthropic's Fable 5 and OpenAI's Daybreak, could transform offensive and defensive cyber capabilities rapidly, exploiting weaknesses like legacy systems and slow patching. The agencies emphasize that older AI models and open-source variants already pose risks, with newer models quickly becoming accessible as development accelerates.

The U.S. government ordered Anthropic to suspend foreign access to its AI models Fable 5 and Mythos 5 over national security concerns related to a potential jailbreak method. Anthropic disabled the models to comply, citing disputes over the severity of the reported vulnerability and noting similar capabilities exist in other public models.

Anthropic released Claude Fable 5, a 'Mythos-class' AI model with safeguards restricting cybersecurity, biology, chemistry, and distillation queries to its Opus 4.8 model. The model exceeds previous Anthropic models in capabilities but is limited in access, with a 'trusted access program' planned for a version without full safeguards.

Anthropic has provided an update on its Project Glasswing initiative. The update was posted on Schneier.com's blog and linked to Hacker News, where it received 21 points but no comments.

UK banks, excluded from Anthropic’s Project Glasswing expansion, are receiving access to OpenAI’s GPT-5.5 Cyber model. While JPMorganChase was the only financial institution granted access to Anthropic’s Mythos Preview, UK banks like HSBC, Lloyds, and Nationwide will use GPT-5.5 instead. The Bank of England’s governor, Andrew Bailey, criticized Anthropic’s exclusion of UK institutions, with some suspecting political motives behind access decisions.

Cisco used Anthropic’s Claude Mythos Preview and OpenAI’s GPT 5.5-Cyber to scan 1.8 billion lines of code in eight weeks, a task that would have taken eight years manually, but did not disclose the number of vulnerabilities found. Anthropic expanded its Project Glasswing partner program to 200 organizations, and Palo Alto Networks reported 26 CVEs discovered using Mythos in a month.

Anthropic has expanded its Claude Mythos preview program to include more partners by inviting approximately 150 additional organizations to join Project Glasswing.

Anthropic is expanding its security vulnerability program, Project Glasswing, and access to Mythos to 150 organizations across 15 countries. The initiative targets critical infrastructure in power, water, healthcare, and communications sectors, aiming to protect systems where a cyberattack could affect 100 million people.

Anthropic is expanding its Project Glasswing program to 150 new organizations across 15 countries, focusing on underrepresented sectors like power, water, and healthcare. The initiative, using its Claude Mythos Preview model, has already identified over 10,000 high- or critical-severity software vulnerabilities. Partners such as Cloudflare and Mozilla reported significant increases in bug discovery rates, but Anthropic emphasizes challenges in addressing these flaws due to human resource limitations.

Anthropic has expanded its Mythos project to 150 additional organizations across more than 15 countries. The company previously released Project Glasswing to approximately 50 partners in April to test the model for cybersecurity flaws.

Anthropic is expanding Project Glasswing, an initiative detailed in their news article. The project has received 14 points and 6 comments on Hacker News.

Anthropic is opening access to Mythos AI for the European Union's ENISA. ENISA's participation in Project Glasswing stems from strong bilateral cooperation between the European Commission and Anthropic.

Anthropic's Project Glasswing, using its Mythos AI model, discovered over 10,000 high- or critical-severity software vulnerabilities in its first month, significantly advancing AI-driven cybersecurity. Partners like Cloudflare and Mozilla reported substantial increases in bug detection, though challenges remain in verifying and patching flaws. The model's success highlights a shift in cybersecurity priorities from discovery to remediation.

Anthropic plans to release Mythos-class AI models to the public once safety safeguards are developed, but acknowledges current risks. The models, highly effective at finding software vulnerabilities, are currently restricted via Project Glasswing due to misuse concerns. Governments and security experts have responded with heightened scrutiny and action.

Anthropic's Mythos has identified over 10,000 vulnerabilities, with Project Glasswing aiding partners in discovering high and critical severity bugs.

Anthropic's Project Glasswing has identified over 10,000 high-severity vulnerabilities in critical software since its launch last month. The initiative, led by AI company Anthropic, aims to enhance cybersecurity by addressing systemic risks.

Anthropic announces an initial update on Project Glasswing, sharing research findings and progress. The project update has generated moderate engagement on Hacker News with 20 points and 9 comments.

Project Glasswing was showcased by Mythos, with an article available on Cloudflare's blog and a comments section on Hacker News. The project has garnered 25 points and currently has no comments. The article discusses cyber frontier models.

OpenAI has launched Daybreak, a cybersecurity initiative that combines large language models with its Codex framework to identify and patch software vulnerabilities. The platform offers three model tiers with different safeguard levels and access controls. Daybreak arrives after Anthropic unveiled Project Glasswing, a cybersecurity-focused AI system.

cURL developer Daniel Stenberg tested Anthropic's Mythos AI model and found it to be primarily a marketing stunt, discovering only one confirmed security vulnerability. The model was expected to find more vulnerabilities, but its findings were mostly false positives or already known issues. Stenberg concludes that Mythos is not a groundbreaking AI model.

The article discusses how advancements in AI, particularly Anthropic's Claude Mythos and Project Glasswing, are accelerating the closure of the 'exploit window' for vulnerabilities. Traditional patching methods are becoming insufficient, necessitating new strategies like Network Detection and Response (NDR) to address emerging threats in a zero-window era.

Anthropic's AI model Claude Mythos can autonomously identify and exploit zero-day vulnerabilities in operating systems and browsers, prompting the company to withhold public release. The model is shared with 40 US-based partners under Project Glasswing to preempt cyber threats, with the UK's AI Security Institute also testing it. British officials warn AI could escalate cyber-attacks, leaving most businesses unprepared.

Anthropic's Project Glasswing, an AI model designed to detect software vulnerabilities, was delayed from public release. The company granted access to Apple, Microsoft, Google, Amazon, and others to identify and patch bugs before malicious actors could exploit them.

Anthropic's Project Glasswing involves over 50 companies testing its Mythos model for security vulnerabilities, but the exact number of discovered issues remains unclear. The initiative aims to identify security flaws in participants' products using the advanced LLM, though results are still uncertain.

OpenAI is expanding its Trusted Access for Cyber program to thousands of individuals and organizations, introducing GPT 5.4 Cyber, a cybersecurity-optimized model. The program aims to enhance accessibility to advanced security tools while enforcing strict verification to prevent misuse, contrasting with Anthropic's Project Glasswing and Claude Mythos.

Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened Wall Street CEOs to address cybersecurity risks from Anthropic's AI model Claude Mythos Preview, which experts warn could enhance cyberattacks. The meeting follows a Pentagon dispute with Anthropic over its technology use in autonomous weapons and surveillance.

Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened Wall Street CEOs for an emergency meeting to address cybersecurity risks from Anthropic's new AI model, Claude Mythos Preview, which can autonomously identify software vulnerabilities. The meeting focused on mitigating threats to the global financial system amid tensions between Anthropic and the Pentagon over its technology use in military applications.

Project Glasswing, a coalition of tech giants, is using Anthropic's Mythos AI to identify and fix vulnerabilities in open source software. The initiative faces scrutiny as The Reg highlights the AI's ability to generate zero-day vulnerabilities, raising concerns about security risks.

Anthropic has launched Project Glasswing, an initiative aimed at preventing AI-powered cyberattacks using artificial intelligence. The project focuses on developing defensive measures against malicious AI activities.