AI security

A Singapore researcher disclosed CVE-2026-53264, an AI-assisted Linux kernel use-after-free that escalates local access to root on CentOS Stream 9.

The Linux Foundation and roughly 20 companies including AWS, Anthropic, Google, IBM and major banks launched Akrites, a coordinated incident-response alliance for open source software, after finding fewer than 5% of recently surfaced vulnerabilities have been patched.

Hugging Face disclosed that an autonomous AI agent system breached its data-processing pipeline over a weekend, harvesting credentials before defenders turned to a self-hosted open-weight model to analyze the attack.

OpenAI says its internal attacker model GPT-Red cut GPT-5.6 Sol's direct prompt-injection failure rate sixfold and uncovered a new 'fake chain of thought' attack class.

Sysdig says JadePuffer used an LLM agent to exploit Langflow flaw CVE-2025-3248, run 600-plus payloads, and encrypt 1,342 Nacos config items.

Socket researchers discovered TrapDoor, a supply chain attack spanning 34 packages and 384+ versions across three registries, with a novel technique that embeds hidden instructions in AI coding assistant config files to trigger credential exfiltration.

After an Alpha Omega-mediated Mythos scan turned up five claimed vulnerabilities, the curl security team triaged the list to one low-severity CVE bound for curl 8.21.0 in late June.

An LLM-powered scanner from security startup DepthFirst flagged a heap buffer overflow that had sat undetected in NGINX's rewrite module for roughly 18 years, prompting F5 to ship coordinated patches on May 13.

Microsoft's new Autonomous Code Security team disclosed MDASH alongside its May 2026 Patch Tuesday, crediting the multi-model agentic scanner with 16 Windows vulnerabilities — four of them critical RCEs — and an 88.45 percent score on CyberGym.

Mozilla released Firefox 150 on April 21, 2026, fixing 271 vulnerabilities surfaced by Anthropic's Claude Mythos Preview in a security sweep Mozilla's CTO calls a turning point for defender-side AI.

CISA added CVE-2026-34197, a 13-year-old remote code execution flaw in Apache ActiveMQ Classic, to its Known Exploited Vulnerabilities catalog on April 16 as Horizon3.ai's Naveen Sunkavally described finding the chain with Anthropic's Claude in about ten minutes.

A coalition of cybersecurity heavyweights warns that AI-driven vulnerability discovery has collapsed mean time-to-exploitation to under one day, urging CISOs to deploy AI agents against their own code immediately.