Topic: Security
407 stories · page 1 of 9
Vulnerabilities in models and agents, prompt injection, data leaks, malicious packages, and the defences against them. Alignment research and the way labs test models before release are covered here too.
- Microsoft Research podcast: Mike Walker on rules, trust and hacker culture
- macOS daemons build a personal graph that scores who is family, 8rank.ai says
- Coldframe open-source honeypot mimics Danfoss and Dixell fridge controllers
- Alex Stamos says open-weight AI models are upending cybersecurity
- Windward tracks Iran-linked ships loitering over Gulf of Oman subsea cables
- Paper shows backdoor attacks that steer multi-turn chats in LLM dialogue systems
- InfoHawk finds 820 fake restaurant sites, some skimming card data
- Firecracker microVM escape achieves root code execution on host
- DeepSeek nearly matches OpenAI's Daybreak in hacking tests, Vals finds
- China reportedly recruited Megaspeed exec to spy on Nvidia's Jensen Huang
- Chainalysis says hackers obtained names, password hashes and live tokens for all its users
- Censys finds 800+ exposed deployments of Artex AI hacking tool
- OpenAI introduces GPT-6 Cyber, a model for finding and fixing vulnerabilities
- Microsoft Paint remote code execution flaw CVE-2026-83844 rated CVSS 8.1
- Decade-old OpenSSH and musl flaw lets attackers SSH into Alpine Linux without a password
- Claude Code reportedly sends user email in User-Agent header
- Possibly Kremlin-linked hackers exploited patched iOS and Android flaws
- PHP2Shell flaw shows AI digging up decades-old PHP bugs
- OpenAI pipeline generates 948 Linux root exploits for $1.2M in API costs
- MCPGuard finds conformance gaps in official MCP Python and TypeScript SDKs
- Binance says hackers used a fake job interview to steal $6.3B in crypto
- Postman app auto-installed plugins into Codex and Claude Code, user reports
- Hackers remotely overheat fridges at European supermarkets, spoiling food
- Google says Android partner Videostrong infected 1.3 million devices
- WIRED finds 400+ AI apps hiding sexual deepfake generators
- Ukraine's Air Alert app hit by cyberattack with fake nuclear warnings; Russia reports its own
- Ubuntu lacks a patch for critical kernel bug CVE-2026-98432, blogger says
- OpenAI raises Security Bug Bounty rewards and adds new bonuses
- Binance sees $3.6bn net outflow after record $6.3bn hack
- Trigger.dev says attacker accessed customer environment variables
- Volta infostealer sends stolen data to LLMs to profile victims, Cisco Talos finds
- RAMBO attack leaks air-gapped computer data via radio signals from RAM
- OpenAI to train models on Daybreak Blue cybersecurity session data from November 29
- Linux kernel CVE-2026-98432: IPv4 TCP use-after-free rated 9.8
- ChatGPT share link could run attacker code in a victim's session
- Northflank reports security incident, tells users to rotate all secrets
- Claude Code extension on Cursor Marketplace carried an SSH backdoor, Phorion says
- Binance loses about $6.3B in crypto heist, tracers suspect North Korea
- Azure API Connections flaws let researcher cross-tenant compromise, netting $200,000
- OpenAI describes layered monitoring to catch frontier model misalignment
- NYT: US believes a Megaspeed executive close to Nvidia's Jensen Huang is a China informant
- Linux kernel CVE-2026-98432: TCP use-after-free rated 9.8 upstream, 5.9 by Red Hat
- Accenture contractor reportedly linked DOD system to personal Claude Code
- Wiz CTO on two critical PHP bugs and AI-speed vulnerability hunting
- MuPDF and Foxit generate weak AES-256 PDF keys, open without a password
- Apple's macOS git and curl still trust compromised eMudhra certificates
- Alpine Linux rebuilds GitLab after attempted CVE-2026-97229 exploit
- TianoCore EDK II UEFI flaw CVE-2026-97631 allows pre-boot code execution
- Tailscale details SSH privilege-escalation flaws, patch targeted for 27 October
- OpenAI says Check Point uses Daybreak to speed up vulnerability research