Loading story…
Reports claim Anthropic’s Claude was used to identify and weaponize CVE-2026-4747, a FreeBSD kernel vulnerability in the RPCSEC_GSS component used for Kerberos authentication in NFS servers. The flaw reportedly allowed an unauthenticated attacker to trigger a stack buffer overflow and gain remote code execution, ultimately obtaining a root shell on unpatched systems. A social media post and a Forbes report both describe the incident as a notable milestone in AI-assisted offensive security, with the advisory crediting Nicholas Carlini using Claude for discovering the bug.
According to the Forbes account, an autonomous AI workflow analyzed the published advisory and produced two working exploits in roughly four hours, moving from vulnerability identification to practical exploitation without the extended manual effort typically required for kernel exploit development. The report argues that this sharply reduces the time between disclosure and weaponization and raises pressure on defenders to accelerate patching, continuous code auditing, and AI-assisted detection and response as similar systems are reportedly being applied to find hundreds of additional high-severity flaws.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
4 events from the most recent confirmed update back to the earliest known activity.
The Forbes article further claimed that the Claude-powered vulnerability research pipeline had since identified 500 additional high-severity vulnerabilities across multiple codebases. No separate dates or affected vendors were provided for those findings.
A public write-up for CVE-2026-4747 described the vulnerable svc_rpc_gss_validate() path in FreeBSD's kgssapi.ko, listed affected FreeBSD versions, and said the issue was fixed by adding a bounds check on oa_length. It also detailed exploitation requirements and a tested multi-round ROP and shellcode chain on FreeBSD 14.4-RELEASE amd64 that could yield a root reverse shell on vulnerable NFS servers.
A subsequent report said an autonomous Claude-based agent analyzed the newly disclosed vulnerability and, within about four hours of the advisory, produced two working exploits that achieved a root shell on unpatched FreeBSD systems. The article framed this as a milestone in AI-driven offensive cyber capability.
A FreeBSD security advisory disclosed CVE-2026-4747 in the RPCSEC_GSS module used for Kerberos authentication in NFS servers. The advisory credited Nicholas Carlini using Anthropic's Claude for discovering the flaw, which could allow unauthenticated remote code execution via a stack buffer overflow.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
7 references tracked. Mallory keeps watching after this page renders.
flyingpenguin.com
Open sourcegithub.com
Open sourceforbes.com
Open sourcegithub.com
Open sourceblog.calif.io
Open sourcesentinelone.com
Open sourcefacebook.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Anthropic disclosed that its **Claude Mythos Preview** model autonomously discovered and, in some cases, exploited vulnerabilities across major software targets, including **Firefox**, **OpenBSD**, **FreeBSD**, **Linux**, **FFmpeg**, cryptography libraries, and a memory-safe virtual machine monitor. The company said the system produced browser exploit chains involving JIT heap sprays and sandbox escapes, chained Linux kernel flaws into local privilege escalation, and developed a fully autonomous exploit for **`CVE-2026-4747`**, a FreeBSD NFS issue that reportedly enables unauthenticated remote root access. Anthropic said most findings remain under coordinated disclosure because fewer than 1% had been patched at the time of publication. Publicly described examples included a **27-year-old OpenBSD SACK denial-of-service bug** that was later addressed in an OpenBSD patch, and a **16-year-old FFmpeg H.264 out-of-bounds write**. Anthropic also said it partnered with Mozilla to improve Firefox security, underscoring that modern AI systems are now capable of meaningful autonomous vulnerability research against widely deployed platforms. To limit immediate misuse, the company withheld most technical details and published **SHA-3 commitments** for future disclosure while urging defenders to accelerate AI-assisted bug hunting, patching, disclosure coordination, and incident-response automation.
Jul 28, 2026
Researchers working with the FreeBSD project disclosed an AI-assisted security audit that found **15 kernel vulnerabilities**, including **five local privilege escalation flaws**, **one `bhyve` guest-to-host escape**, and multiple memory disclosure and denial-of-service issues. The effort was coordinated directly with FreeBSD maintainers, with the researchers saying many reports moved from disclosure to fix within days and that the goal was to surface serious, actionable bugs rather than inflate vulnerability counts. The researchers also publicly released exploits and AI-generated writeups for three verified local privilege escalation vulnerabilities: **`CVE-2026-45250`**, **`CVE-2026-45253`**, and **`CVE-2026-45251`**. They said their workflow focused on reporting only high- and critical-severity issues, keeping submissions concise, and optionally proposing patches while maintaining direct communication with maintainers; they added that similar AI-assisted auditing is underway for other open-source infrastructure projects.
Jul 24, 2026
**Check Point Research** disclosed multiple critical vulnerabilities in Anthropic’s *Claude Code* AI coding assistant that could allow **remote code execution** and **credential theft** when a developer clones and opens an **untrusted repository**. The reported attack path abuses repository-controlled configuration and automation features (including **Hooks**, **MCP servers**, and **environment variables**) to trigger hidden shell command execution and to exfiltrate **Anthropic API credentials**, potentially enabling a pivot from a developer workstation into broader enterprise environments where Claude-related workflows and shared resources are accessible. The issues include consent-bypass and command-execution weaknesses tracked under **CVE-2025-59536** (covering closely related flaws involving repository configuration executing commands without adequate user consent) and an API credential exposure issue tracked as **CVE-2026-21852**, which affected *Claude Code* versions prior to **2.0.65** and enabled API key theft via malicious project configurations. Anthropic has **patched** the vulnerabilities and advised users to update to the latest version, while indicating additional hardening measures are planned to reduce supply-chain risk from malicious commits and repository-level configuration abuse.
Jun 29, 2026