SECURITY · OpenAISECURITY
Security & AI
AI security, prompt injection and threats.
SECURITY · OpenAI
SECURITY · AnthropicAnthropic cuts its vulnerability disclosure window from 90 to 30 days for serious flaws
2 MIN READ · NOV 17, 2026
SECURITY · CormaCorma raises $125 million at a $1 billion valuation for defensive cybersecurity AI models
2 MIN READ · NOV 17, 2026Google says an attacker framework turned security patches into working exploits with no human in the loop
2 MIN READ · NOV 16, 2026
SECURITY · North KoreaUS disrupts North Korea's 'FaceStealer' fake video-call scheme, tied to more than $400 million in crypto theft
2 MIN READ · NOV 16, 2026
SECURITY · CounterintelligenceAI labs ramp up counterintelligence hiring as espionage fears grow, with Anthropic's hires more than doubling
2 MIN READ · NOV 15, 2026
SECURITY · Linux FoundationLinux Foundation launches Agent Name Service to give AI agents DNS-based identities
2 MIN READ · NOV 15, 2026
SECURITY · AmazonAmazon decoys posing as LLM APIs capture ARTEX agents hunting AI keys and attempting deepfake KYC fraud
2 MIN READ · NOV 14, 2026
SECURITY · IBMIBM previews Granite 4.1 LLMON models that use a dedicated markup to resist prompt injection
2 MIN READ · NOV 14, 2026
SECURITY · OpenAIOpenAI launches Daybreak Red Team and Defense Factory to find, validate and patch vulnerabilities with agents
2 MIN READ · NOV 14, 2026
SECURITY · Microsoft ResearchMicrosoft researchers show that editing fewer than ten neurons can plant a backdoor in an AI agent
2 MIN READ · NOV 14, 2026
SECURITY · HelsingHelsing unveils Clarity, an AI agent platform that prepares military targets for human approval
2 MIN READ · NOV 13, 2026
SECURITY · Fireworks AIFireworks AI breach via cloud provider exposes Notion AI chat transcripts
2 MIN READ · NOV 13, 2026
SECURITY · Claude CodeApiiro finds Claude Code's "don't ask again" approvals left secrets in 1,122 public GitHub files
2 MIN READ · NOV 8, 2026
SECURITY · CybersecurityFree Chinese open-weight model comes close to OpenAI's restricted cyber model in simulated network attacks
2 MIN READ · NOV 8, 2026
SECURITY · OpenSSHOpenSSH 10.6p2 fixes a long-password login bypass on musl Linux, found by a researcher using AI agents
2 MIN READ · NOV 8, 2026
SECURITY · OpenAIOpenAI signs testing partnership with Hacktron, the firm that breached its systems through a bug bounty
2 MIN READ · NOV 8, 2026
SECURITY · AnthropicClaude deletes a law professor's syllabus and lecture notes during a page-renumbering task
2 MIN READ · NOV 8, 2026
SECURITY · DeepfakesUS prosecutors charge three men over alleged $129 million scheme using AI to impersonate CEOs and lawyers
2 MIN READ · NOV 7, 2026
SECURITY · CensysCensys counts about 1,465 exposed ARTEX AI hacking consoles on the open internet
2 MIN READ · NOV 7, 2026
SECURITY · BotpressBotpress discloses breach that exposed bot configurations and stored credentials in 16 customer workspaces
2 MIN READ · NOV 7, 2026
SECURITY · Check PointCheck Point: ARTEX agents stole iOS signing certificates and game logins with little human input
2 MIN READ · NOV 7, 2026
SECURITY · Frontier Model ForumFrontier Model Forum proposes a shared scale for rating cyber jailbreaks
2 MIN READ · NOV 7, 2026
SECURITY · Google DeepMindGoogle DeepMind says AI agents helped find two new weaknesses in legacy 1024-bit RSA
2 MIN READ · NOV 7, 2026
SECURITY · Collinear AICollinear AI releases CWE-bench v2, 100 runnable tasks testing whether coding agents can patch vulnerabilities
2 MIN READ · NOV 6, 2026
SECURITY · AnthropicAnthropic makes Claude Enterprise inference hooks generally available and opens a security center beta
2 MIN READ · NOV 6, 2026
SECURITY · WorkOSWorkOS says attacker used the ARTEX AI framework to break into an internal sales app
2 MIN READ · NOV 5, 2026
SECURITY · OpenAIOpenAI releases GPT-6 Cyber to a small group of Daybreak Red customers
2 MIN READ · NOV 5, 2026
SECURITY · Palo Alto NetworksPalo Alto Networks reportedly in talks to buy AI agent identity startup Clutch Security for about $500 million
2 MIN READ · NOV 5, 2026
SECURITY · Check PointCheck Point finds 'prompt adware' in a Codex jailbreak bundle that pushes a paid LLM gateway
2 MIN READ · NOV 5, 2026
SECURITY · OpenAIOpenAI says an agent pipeline produced exploits for 948 Linux privilege-escalation flaws in one run
2 MIN READ · NOV 5, 2026
SECURITY · BinanceBinance says hackers used a fake job interview and an internal AI tool in its $6.3 billion theft
2 MIN READ · NOV 5, 2026
SECURITY · XenXen patches guest-to-host memory flaws that Anthropic found using Claude (XSA-523)
2 MIN READ · NOV 5, 2026
SECURITY · LLMjackingGoogle and Mandiant warn that LLMjacking can run up cloud AI bills of $100,000 a day
2 MIN READ · NOV 4, 2026
SECURITY · HBMSmugglers reportedly bundle Korean HBM with throwaway processors to slip it into China
2 MIN READ · NOV 4, 2026
SECURITY · AnthropicAnthropic's FlashProve cryptographically proves a response came from committed model weights
2 MIN READ · NOV 4, 2026
SECURITY · CrowdStrikeCrowdStrike to acquire AI agent security startups Fiddler and Dreadnode
2 MIN READ · NOV 4, 2026
SECURITY · OpenAIOpenAI raises bug bounty payouts to as much as $172,000 and plans a new security sprint with Cathedral
2 MIN READ · NOV 3, 2026
SECURITY · NCSCUK's NCSC judges AI-powered ransomware on critical infrastructure 'highly likely' by 2028
2 MIN READ · NOV 3, 2026
SECURITY · OpenAIOpenAI says activation probes have screened every assistant completion since GPT-5.6, token by token
2 MIN READ · NOV 3, 2026
SECURITY · NorthflankNorthflank breach: customers' AI API keys used without authorization
2 MIN READ · NOV 2, 2026
SECURITY · AI agentsStudy: AI agents raised their own crypto wallet spending caps by reading emailed approval codes
2 MIN READ · NOV 2, 2026
SECURITY · ChainalysisChainalysis says an intruder using 'offensive agentic techniques' reached customer names and emails
2 MIN READ · NOV 2, 2026
SECURITY · OpenAIOpenAI tells Daybreak Blue users it will train on their cyber sessions from November 29
2 MIN READ · NOV 1, 2026
SECURITY · AnthropicAnthropic extends Claude's text watermark to older models, with a sampling caveat on two
2 MIN READ · NOV 1, 2026
SECURITY · OpenAIOpenAI fixes ChatGPT shared-link flaw that let attacker code read private chats
2 MIN READ · NOV 1, 2026
SECURITY · Cisco TalosCisco Talos details VOLTA, an infostealer that asks an LLM which victims are worth pursuing
2 MIN READ · NOV 1, 2026
SECURITY · AnthropicAnthropic had Claude craft training data that slips past its own monitor and makes models misaligned
2 MIN READ · OCT 31, 2026
SECURITY · QwenNudging two neurons turns Qwen3.5-4B hostile, as a community catalog of steering recipes grows
2 MIN READ · OCT 31, 2026
SECURITY · AnthropicAnthropic trains Claude to flag misuse in its own conversations, catching 94% of high-priority cases
2 MIN READ · OCT 31, 2026
SECURITY · CrowdStrikeCrowdStrike adds agent-planning attacks to its prompt injection taxonomy, now 222 techniques
2 MIN READ · OCT 31, 2026
SECURITY · Claude CodeTrojanized Claude Code extension on Cursor's marketplace planted SSH backdoors on developer machines
2 MIN READ · OCT 30, 2026
SECURITY · MicrosoftMicrosoft says eight North Korean hacking groups now use AI across their operations, still under human direction
2 MIN READ · OCT 30, 2026
SECURITY · CrowdStrikeCrowdStrike says China-nexus VAPOR PANDA used the ARTEX agent framework to target more than 200 organizations
2 MIN READ · OCT 30, 2026
SECURITY · EMBRACEEMBRACE research finds browser security gaps in Cursor, Copilot, Codex, VS Code and Claude Desktop
2 MIN READ · OCT 30, 2026
SECURITY · PentagonAccenture staffer's personal Claude agent reportedly ran on a Pentagon system for two months
2 MIN READ · OCT 30, 2026
SECURITY · EPICSCausal Security's AI agent found four EPICS Base flaws, one critical, in under four hours
2 MIN READ · OCT 29, 2026
SECURITY · MCPCritical flaw in Google's MCP Toolbox let attackers read and overwrite server files
2 MIN READ · OCT 29, 2026
SECURITY · CloudflareCloudflare says a KV cache flaw leaked other customers' data in 155 GLM-5.3 responses
2 MIN READ · OCT 28, 2026
SECURITY · CohereCohere says Alibaba's Qwen3.8 tried an unprompted upload to a Beijing server during a sandboxed test
2 MIN READ · OCT 28, 2026
SECURITY · AnthropicAnthropic says an unreleased Claude model built working exploit chains for Chrome, Safari, Android, Windows and Linux
2 MIN READ · OCT 28, 2026
SECURITY · CERT-InIndia's CERT-In warns AI agents under evaluation accessed live systems without authorization
2 MIN READ · OCT 28, 2026
SECURITY · OWASPPrompt injection tops the OWASP list of LLM application risks
1 MIN READ · OCT 28, 2026
SECURITY · CopilotA trailing dot let Copilot in Visual Studio rewrite its own guardrails, enabling code execution
2 MIN READ · OCT 27, 2026NEWSLETTER · FREE
Your AI briefing, filtered by your interests
Every morning at 7 a.m. your local time, only the topics you choose. In English, French or Spanish.
Choose your topicsOne-click unsubscribe. We never sell your data.