Artificial Intelligence (AI)
Anthropic's Fable 5 launches with “cyber safeguards” baked in — a framing that would, days later, become the center of a precedent-setting US export-control action.
Artificial Intelligence (AI)
Half a century after Brunner imagined it, researchers publish a prototype — and the defender community gets a new detection-research agenda rather than an immediate operational threat.
Vulnerabilities
A use-after-free in the Linux kernel's nf_tables code — patched in February, exploited publicly in June — shows how a single misplaced character in a critical subsystem becomes the keystone of a privilege-escalation chain.
Artificial Intelligence (AI)
Bright Data, formerly Luminati, runs the largest residential proxy network in the world — and a researcher has now mapped how consumer apps and always-on smart TVs become its silent infrastructure.
Artificial Intelligence (AI)
OpenAI's Lockdown Mode is the first big consumer-facing prompt-injection defense from a frontier AI lab — but the company itself concedes the feature reduces, rather than eliminates, the risk.
Nation-State Cyber Threats
ReliaQuest disclosed OP-512, a previously unreported, China-linked espionage cluster that plants a custom three-web-shell framework on Microsoft IIS servers — the fourth such group to target IIS in a year. For anyone running IIS, it is a prompt to go hunting.
Supply Chain Attack
Three disclosures this cycle share one thesis: attackers borrowing the trust of legitimate channels. A Rust-written npm worm (IronWorm), a cryptominer slipped into Hola Browser, and a Magecart skimmer hosted inside Stripe each hide in traffic defenders are inclined to allow.
Artificial Intelligence (AI)
Two Mythos threads landed this cycle: TechCrunch reports the NSA is said to be readying Anthropic's Mythos for cyber operations despite a federal restriction, while Anthropic published an analysis of 832 accounts banned for malicious cyber activity, mapped to MITRE ATT&CK.
Cybersecurity 101
Digital forensics is the disciplined reconstruction of what happened in a cyber incident. A complete guide to its branches, the six-stage process, chain of custody, DFIR, and the 2026 challenges of cloud, encryption, and anti-forensics.
Vulnerabilities
Two vulnerabilities disclosed this cycle were found by AI tooling: HTTP/2 Bomb (CVE-2026-49975), a remote DoS that crashes NGINX, Apache, IIS, Envoy and Cloudflare Pingora in default config, and CVE-2026-23479, a two-year-old authenticated RCE in Redis.
Vulnerabilities
Researcher Ammar Askar disclosed a one-click attack via VS Code's GitHub.dev that steals a GitHub OAuth token with read-write access to private repos. He published the PoC with about an hour's notice, blaming Microsoft's disclosure process.
Artificial Intelligence (AI)
SafeBreach researchers found that a single poisoned notification — from WhatsApp, Slack or SMS — could hijack Google Gemini's voice assistant on Android with no malicious app installed, reaching smart-home controls and poisoning the assistant's long-term memory. Google has patched it.