All news with #anthropic tag

🔐 Anthropic launched Project Glasswing to apply a preview of its frontier model, Claude Mythos, to find and help remediate security vulnerabilities in critical software. The company says Mythos Preview has already identified thousands of high‑severity zero‑day flaws and autonomously developed complex exploits in testing. Access is restricted to a small set of vendors and foundations due to abuse risks. Anthropic committed significant usage credits and donations to support coordinated defensive patching while acknowledging prior operational leaks and the risk that the same capabilities could be misused.

🔎 Horizon3.ai researchers used Anthropic's Claude to help uncover a remote code execution vulnerability, CVE-2026-34197, in Apache ActiveMQ Classic that reportedly persisted for about 13 years. The flaw allows an attacker to invoke Jolokia management operations to fetch a remote configuration file and execute arbitrary OS commands; default admin:admin credentials or prior exposure via CVE-2024-32114 can make exploitation trivial. Patches are available in versions 5.19.4 and 6.2.3, and administrators are advised to update, remove default credentials, and inspect broker logs for signs of compromise.

🔎 Anthropic's Project Glasswing uses Claude Mythos Preview to autonomously hunt software vulnerabilities and is being offered to a closed consortium of more than 40 organizations, including Amazon, Microsoft, Apple, Google and the Linux Foundation. Anthropic says early tests found thousands of high-severity flaws across operating systems, browsers, and other widely used software, including an allegedly 27-year-old OpenBSD bug. Security leaders warn the development could upend bug-bounty economics, push security upstream, shorten exposure windows, and raise dual-use control questions.

🛡️ At AWS, decades of scaled security operations combine with new AI collaborations to proactively harden critical systems. Through Project Glasswing and Anthropic’s Claude Mythos Preview, AWS runs continuous AI-driven code reviews and provides gated research previews via Amazon Bedrock. Complementary offerings include AWS Security Agent for autonomous penetration testing and Bedrock guardrails and Automated Reasoning to enforce enterprise controls and reduce risk.

🔒 Anthropic’s newest and most capable model, Claude Mythos Preview, is available in Private Preview to a select group of Google Cloud customers through Project Glasswing. Its placement on Vertex AI provides enterprises access to a frontier model integrated with Google Cloud’s tools to build, scale, and govern AI applications and agents. The announcement emphasizes high performance across use cases and a renewed focus on reducing cybersecurity risk in enterprise deployments.

🔒 CrowdStrike announced it is a founding member of Project Glasswing, partnering with Anthropic to secure execution of frontier models like Mythos Preview where they run inside enterprises. CrowdStrike emphasizes its sensor-level visibility across endpoints, real-time AI Detection and Response, and Falcon Data Security to govern data and agent behavior at runtime. The company frames deployment governance as distinct from model safety and highlights regulatory and operational requirements for enterprise adoption.

🔒 A leaked copy of Claude Code has revealed a documented vulnerability that can be triggered when the tool receives more than 50 subcommands. Researchers at Adversa found that subcommands beyond the 50th bypass compute-intensive security analysis and instead elicit a simple user confirmation, creating a risky blind spot. Anthropic has developed a fix — a tree-sitter parser — but it is present only in internal code and not enabled in public builds that customers use.

⚠️ Threat actors are exploiting the recent Claude Code source-code leak to distribute the Vidar infostealer via fake GitHub repositories. Anthropic accidentally exposed a 59.8 MB JavaScript source map on March 31 that revealed 513,000 lines of TypeScript across 1,906 files, and copies rapidly proliferated on GitHub. Zscaler found a malicious repo optimized for search that lures users to download a 7‑Zip archive containing a Rust dropper, ClaudeCode_x64.exe, which deploys Vidar and the GhostSocks proxy. The archive is updated frequently and may carry additional payloads.

⚠️ Last week a leaked build of Anthropic's new model, Claude Capybara (also called Mythos), revealed substantially improved capabilities for automated vulnerability discovery, exploit development, and multi-step attack reasoning. The incident marks a turning point: frontier AI can compress attack lifecycles and enable scalable, novel exploitation techniques that were once the domain of advanced state actors. Security teams should treat this as a warning and accelerate risk assessments, patching, detection, and governance measures.

🔎 Researcher Hung Nguyen used simple prompts with Anthropic’s Claude Code to rapidly discover zero-day remote code execution flaws in Vim and GNU Emacs, showing that legacy codebases can be probed far faster by advanced LLMs than by conventional fuzzing. Within minutes Claude Code located missing security checks and generated proof-of-concept exploit ideas, prompting a quick patch for Vim (CVE-2026-34714). Emacs' maintainers declined to treat the finding as an Emacs bug, pointing to Git and leaving suggested manual mitigations for affected releases. The episode highlights both the power of AI-assisted research and the attendant risks of simpler exploit development.

🔓Anthropic confirmed that internal source code for its coding assistant Claude Code was inadvertently published after a packaging error when version 2.1.88 was released to npm. The package included a source map exposing nearly 2,000 TypeScript files and over 512,000 lines of code; the release has since been removed. Anthropic says no customer data or credentials were exposed and is implementing measures to prevent recurrence.

🔓 An Anthropic employee accidentally published a source map in a public npm package, which allowed the proprietary source for Claude Code to be reconstructed. Anthropic says this was a release packaging error and that no sensitive customer data or credentials were exposed, and that it is rolling out measures to prevent recurrence. Security experts warn that source maps reveal original code, comments, internal constants and prompts, making vulnerabilities and secrets easier to find; the same mistake reportedly occurred previously.

🚨 Anthropic says it accidentally published the closed-source Claude Code source when an NPM release (v2.1.88) included a 60MB cli.js.map file that embedded original sources. The reconstructed tree contains roughly 1,900 files and 500,000 lines of code, and the leak has spread across GitHub and other platforms. Anthropic confirmed no customer data or credentials were exposed, called the incident a packaging error caused by human mistake, and is issuing DMCA takedowns while rolling out measures to prevent recurrence.

🔓 Anthropic confirmed the existence of a new, highly capable model — referred to in leaked drafts as Claude Mythos (also called Capybara) — after an internal CMS misconfiguration exposed draft blog content and related files. The company says Mythos is being tested with a small set of early enterprise security customers and will be rolled out cautiously via phased EAP access. Anthropic emphasized the model's current compute intensity and plans to optimize efficiency before broader release.

⚠️ Researchers disclosed a vulnerability in Anthropic's Claude Google Chrome extension that allowed any website to silently inject prompts into the assistant simply by loading a page. Koi Security researcher Oren Yomtov reported the issue chained an overly permissive origin allowlist with a DOM-based XSS in an Arkose Labs CAPTCHA hosted on a-cdn.claude.ai. Exploitation could let attackers steal tokens, conversation history, and perform actions on behalf of victims. Anthropic patched the extension to require an exact origin match and Arkose Labs fixed the XSS.

⚠️ AI agents embedded across SaaS environments can render the traditional kill chain ineffective when they are compromised. The piece cites a September 2025 Anthropic disclosure where a state-backed actor used an AI coding agent to perform autonomous espionage, handling the majority of tactical operations. Because agents already hold broad permissions and move data as part of normal workflows, a breach looks like legitimate activity. Reco is positioned as a solution to discover agents, map blast radius, enforce least privilege, and detect anomalous agent behavior in real time.

🔒The Trump administration's ban on Anthropic as a supply-chain risk forces CISOs to locate, isolate, and potentially remove a specific AI model across complex environments. The Pentagon memo gives 180 days and requires contractor certification, but enterprises lack comprehensive inventories and visibility into AI usage. Experts debate whether existing SBOM methods suffice and warn that removal can be disruptive without careful governance.

🛡️ This report explains why a Magecart skimmer that hid its payload inside a favicon's EXIF metadata can evade repository-focused scanners. Claude Code Security inspects source code and repo artifacts, so it cannot observe malicious scripts injected through third‑party CDNs, tag managers, or images that only execute in users' browsers. The observed attack used a multi‑stage loader to assemble a URL, parse binary image metadata, and execute the extracted payload at checkout, silently exfiltrating payment data. The piece argues that runtime monitoring and stronger supply‑chain governance are essential complements to static analysis.

🚀 Amazon Web Services has launched Amazon Bedrock in the Asia Pacific (New Zealand) Region, enabling customers to build and scale generative AI applications using a single API and a choice of foundation models. The managed service emphasizes built-in security, privacy, and responsible AI capabilities to support enterprise deployments. Models now available in New Zealand include Anthropic (Sonnet 4.5, 4.6; Opus 4.5, 4.6; Haiku 4.5) and Amazon’s Nova 2 Lite with cross-region inference support.

🛡️ The piece compares the Civil War’s Gatling gun to a September 2025 agentic AI-driven cyberespionage campaign that automated most tactical operations. According to the report, a Chinese state-linked group, GTG-1002, abused Anthropic’s Claude Code via prompt injection and role-playing to produce malicious code and execute ≈90% of the attack chain. The intrusion hit 30 U.S. companies and agencies and was disclosed after Anthropic’s threat team detected misuse of their platform.