American Biotech — The Future Can’t Be Won Without It.·RSI is AI Building AI. What DC needs to know about it. ·Keeping the Human in Medicine: Two AI CoLab Researchers on What Machines Can’t Replace·Someone Is Hacking the Freezers. Pay Attention.·Thinking of Running in 2028? Start Here·What the Lame Duck Session Has to Do With the Ultimate 21st-Century Power Tool: Robots·
American Biotech — The Future Can’t Be Won Without It.·RSI is AI Building AI. What DC needs to know about it. ·Keeping the Human in Medicine: Two AI CoLab Researchers on What Machines Can’t Replace·Someone Is Hacking the Freezers. Pay Attention.·Thinking of Running in 2028? Start Here·What the Lame Duck Session Has to Do With the Ultimate 21st-Century Power Tool: Robots·
One card per story, every outlet that covered it55›‹
3d agoThe Register · research by F5AI found more critical bugs in another widely used technology running 225M websitesRapid7 discloses CVE-2026-123007, a critical NGINX HTTP/2 integer underflow that can lead to unauthenticated remote code execution under certain conditions, with F5 fixes issued in NGINX 1.31.9 and 1.30.7.7 articles · BleepingComputer, TechRadar
YesterdayANS Nuclear NewswireComments close soon on PRM for fuel reprocessing plantsA University of Tennessee–Knoxville professor's petition asking the NRC to subject fuel reprocessing plants to Subpart E material control and accounting rules is open for public comment under docket NRC-2026-4456.1 outletEnergy21 linked
3d agoThe Register · research by OX SecurityMicrosoft's official Python agent framework hijacked to distribute malwareBleepingComputer reports that attackers compromised Microsoft's official Agent Framework PyPI package, publishing two malicious versions (1.26.1 and 1.26.2) that steal cloud and developer credentials on Python startup.4 articles · BleepingComputer, TechRadar
YesterdayPalo Alto Unit 42Dialing for Data: Inside the Com’s High-Profile HeistsUnit 42 publishes a research report detailing Com-affiliated vishing, AiTM phishing and data-extortion campaigns, including two newly identified activity clusters (Vortex and Cinder) and a malicious sideloaded browser extension tied to CL-CRI-1192.1 outletCyber5 linked
2d agoWiz ResearchPreparing for Post-Quantum Cryptography | Wiz BlogWiz publishes a guide urging organizations to inventory quantum-vulnerable cryptography and prioritize PQC key exchange upgrades, and announces its Wiz for Post-Quantum Cryptography Security Framework for customer discovery.1 outletQuantumCyber15 linked
2d agoWiz ResearchWiz is on the FedRAMP Marketplace | Wiz BlogWiz announces it is listed as "in process" on the FedRAMP Marketplace as it works toward FedRAMP Moderate authorization for its CNAPP, and hires Dean Scontras as Senior Director of Public Sector.1 outletCyber9 linked
8d agoAnthropicU.S. elections readinessAnthropic's policy page logs the revision history of its Responsible Scaling Policy, including the August 2026 Risk Report and the July 8, 2026 version 3.4 update.12 articles · 4 outlets · TechTarget
4d agoRapid7Open by Default, Already Occupied: Threat Actors are Renting your ComputeRapid7 analysis finds that more than half of exposed Ollama servers in a 2026 scan show signs of prior third-party use, including callback model names, IP-stamped tags, and encoded file-path probes, and that many scanner findings are likely deception infrastructure.1 outletCyber6 linked
11d agoCyberScoop · research by MicrosoftMicrosoft fixes over 1,000 vulnerabilities in November Patch TuesdayMicrosoft's November 2026 Patch Tuesday release fixes 1,094 flaws, including an exploited Outlook spoofing zero-day (CVE-2026-124194) that can leak NTLMv2 hashes.11 articles · 9 outlets · Daily Mail, Fox News
7d agoBleepingComputer · research by CursorCursor confirms security breach led to malicious Claude Code extensionCursor disclosed that a breach traced to stolen Fireworks AI credentials let attackers push trojanized Claude Code extension versions to its marketplace, exposing up to 5,000 developers to attempted SSH backdoors.3 articles
8d agoAnthropicLLMs and bioriskAnthropic introduces Natural Language Autoencoders (NLAs), a method that converts Claude's internal activations into readable natural-language explanations, validated through reconstruction training and safety-audit experiments.13 articles · 2 outlets · BleepingComputer
6d agoWiz Research · research by NordStellarLiteLLM Compromised, Steals Credentials | Wiz BlogWiz Research maps the cloud, code and AI secrets that infostealer malware harvests from developer endpoints and details how attackers turn them into cloud access, using NordStellar telemetry to rank stealer families and secret types.2 articles
6d agoWiz ResearchBuilding an AI Detection Engine for Agent Intent | Wiz BlogWiz outlines its AI agent intent-detection pipeline, which uses model I/O telemetry, layered LLM analysis, and runtime correlation, validated against a simulated benchmark and referencing the OpenAI–Hugging Face incident.2 articles
9d agoBleepingComputer · research by AikidoHackers compromise 4,800 GitHub accounts to steal cloud credentialsThreat actor behind the 'wfx-loot' campaign used about 4,800 compromised GitHub accounts to inject credential-stealing workflows into tens of thousands of repositories and stole CI and cloud secrets from at least 1,835 targets.2 articles
9d agoTechTargetMicrosoft CISO urges Artex regulation, but experts say it won’t workMicrosoft's global CISO Igor Tsyganskiy urged policymakers to regulate release and use of Artex, an open-source LLM-driven pen testing framework, while security experts argue regulation would hamper defenders without curbing criminal use.1 outletCyberAI32 linked
8d agoSemafor · research by Tech Oversight ProjectTrump’s AI task force discusses cyber risksFormer Treasury Secretary Hank Paulson warned at a Semafor event that markets are underestimating AI risk, with Rishi Sunak largely agreeing while criticizing the EU AI Act as the wrong approach.27 articles · 3 outlets · The Hill
11d agoComputerworldThe bad news for Microsoft in the Blue TsunamiComputerworld columnist argues that Democratic midterm wins will bring investigations, data center crackdowns, AI regulation and ICE budget cuts that threaten Microsoft's business.1 outletAI5 linked
12d agoArs Technica Space · research by GreyNoiseVulnerability that lets hackers SSH into Linux devices is under active exploitationAttackers are actively exploiting the SSHole flaw (CVE-2026-123158) in OpenSSH on musl-based Linux systems to bypass password authentication, with GreyNoise recording successful in-the-wild exploitation chains including attempts at persistence and a TAC Agent implant.3 articles · BleepingComputer, Ars Technica
11d agoWiz Research · research by WizHacking GitHub with a Semicolon & Claude with Sagi Tzadik | WizWiz podcast episode in which researcher Sagi Tzadik explains how a semicolon in GitHub's git push option handling enabled a critical RCE, found using Claude paired with reverse engineering tooling.1 outletCyber17 linked
11d agoWiz Research · research by WizOperation Stormbreaker case study | WizWiz published a vendor case study describing how the Marine Corps Community Services (MCCS) used Wiz in its Operation StormBreaker initiative to cut Authorization to Operate timelines from about 18 months to 30 days.1 outletCyber13 linked
11d agoDark ReadingAI Coding Agents Leak Thousands of Secrets to GitHubDark Reading reports that a Claude Code command can inadvertently commit credentials and other sensitive files to public GitHub repositories, exposing organizations to security gaps.1 outletCyber7 linked
12d agoDark ReadingNIST: Solo Shouldn't Mean Less SecureNIST has published Cybersecurity White Paper 50, a plain-language cybersecurity guide for non-employer "solopreneur" firms, which experts say is a useful baseline but has gaps around domain security, email authentication, break-glass access, and agentic AI.1 outletCyber5 linked