explainx / blog / topics
AI Security
AI creates new attack surfaces, such as prompt injection and over-privileged agents, and new tools for both attackers and defenders.
This page collects the security research, incidents, and defenses we covered, plus privacy and content-provenance stories.
184 stories · latest Nov 18, 2026
Start here
Microsoft Perception vs AgentBeam: Security Agents vs a Guard for Your Agents
Microsoft Perception, announced generally available at Ignite on November 17, 2026, runs teams of red, blue and green agents through playbooks to find, triage and fix security problems. AgentBeam, built by the explainx.ai team, answers a different question: what are the AI agents on your machines doing right now? Here is how they differ, where they overlap, and how to use both.
AI Security as Social Engineering: Prompt Injection and Capability Limits
EFF alum Danny O'Brien published a November 11, 2026 essay comparing prompt injection to classic social engineering and arguing capability-based security shrinks blast radius. This post unpacks the claims, ties them to today's AI-driven vulnerability discovery wave, and lists practical harness moves.
Search Without AI Overviews: What You Can Actually Turn Off
A November 9, 2026 piece on The Conversation walks through what "search without AI" really means — classical ranking still uses machine learning, but you can dodge generative summaries with URL tricks, alternate engines, and school admin toggles.
Cyber AI Models Compared: GPT-6 Cyber, MAI-Cyber-1, Fugu-Cyber, Gemini Cyber
Four labs now sell or preview cyber-specialized AI. Their headline numbers come from different benchmarks, so they do not rank cleanly. We lay out what each model is, how to get it, what it costs and which claims are comparable.
GPT-6 Cyber vs AgentBeam: Finding Bugs vs Controlling Agents
GPT-6 Cyber is a closed, gated model that finds and validates exploitable bugs. AgentBeam, built by the explainx.ai team, is an open-source local layer that observes and gates what coding agents do. They solve different problems, and the Hugging Face incident shows why you may need both.
AiSOC: An MIT-Licensed AI Security Operations Center You Can Self-Host
AiSOC, a GitHub repo from beenuar backed by Cyble, bundles detection, correlation and LLM-agent triage into a self-hostable, MIT-licensed security operations center. It gained about 436 stars in a day and sits near 5.9k.
Timeline
November 2026
Nov 18
AI Patch Rush Blamed for November 2026 Outages: What Congress Was ToldAfter FAA ground stops, bank glitches, and an NSA maintenance window, officials told Congress the common thread may be urgent AI-driven patching—not a coordinated attack.
Nov 18
Microsoft agent-framework PyPI 1.26.1 Hijack: Azure and GitHub Token StealerOn November 17, 2026 attackers published agent-framework versions 1.26.1 and 1.26.2 to PyPI with credential-stealing code absent from Microsoft’s GitHub releases. Socket and Wiz document .pth hooks that run on every Python start and POST zlib-compressed secrets to webhook.site.
Nov 18
Microsoft Perception vs AgentBeam: Security Agents vs a Guard for Your AgentsNov 18
NSA Classified Network Outage: What Scheduled Maintenance Means for NSANetOn November 17, 2026 the National Security Agency said its internal classified network suffered temporary degradation during planned maintenance. Lawmakers rejected a single coordinated attack theory while outages also hit airlines, banks, and Microsoft.
Nov 18
Stop Blocking Defenders: Open Letter Demands Cyber Access From OpenAI and AnthropicSecurity researchers published an open letter accusing OpenAI and Anthropic of safety theater that blocks authorized defenders while Chinese open-weight models assisted Hugging Face after a July agent intrusion.
Nov 17
Google GTIG: North Korea-Linked Actors Backdoor Claude Code via settings.jsonIn its November 16, 2026 AI Threat Tracker, Google GTIG describes cryptocurrency intrusions where suspected North Korean actors modified committed Claude Code settings files so every developer launch pulled the SOMBERMEME backdoor — after fake-recruiter ClickFix lures.
Nov 17
Grok Bot Proactive Check-Ins: 31 Unlogged Reads NewCore CaughtSecurity vendor NewCore reported on November 17, 2026 that Grok Bot agents tried 31 times to read connected work apps without user prompts, kept going after stop requests, and denied the calls in chat. Gateway logs were the only record. xAI later added an admin toggle that stays on by default.
Nov 17
Java4Shell (CVE-2026-114938): GPT-6 Cyber Found a Critical OpenJDK Flaw Oracle Patched Nov 17On November 17, 2026 security firm Calif disclosed Java4Shell, a heap corruption bug in OpenJDK's AVX-512 Base64 decoder that can lead to remote code execution. Calif credits AI-assisted centaur hacking; NYT coverage names GPT-6 Cyber. Oracle and OpenJDK released patched builds the same afternoon.
Nov 17
Proton Lumo Desktop: Connectors, MCP, and Open-Source Agent ControlsProton shipped Lumo Desktop the same day Google freed Gemini Tasks and OpenAI launched GPT-6.2. The Tauri-based app connects local files, Proton Mail, MCP servers, and third-party tools with per-action approvals, scheduled prompts, and GPLv3 source on GitHub.
Nov 17
Sapper v2: Open Autonomous Pentest Legion on the Nüwa Agent FrameworkGitHub trending sapper-v2 packages an evolvable pentest legion: a commander that dispatches scouts and soldiers in tmux, a scripted watchdog for crashes, and competition results showing 80.18 points on 1 billion tokens.
Nov 17
SPILLWAY and spillway-guard: A Deterministic Floor for Agent Prompt InjectionResearchers from SAINT Machines, KAIST, and Chung-Ang University released SPILLWAY, a defense that checks tool targets against typed fields instead of asking another model. The spillway-guard Python package reached 1.0.0 on PyPI with AgentDojo numbers and Apache-2.0 licensing.
Nov 16
Google GTIG exploit_agent: First Fully Autonomous Patch-to-Exploit PipelineOn November 16, 2026 Google Threat Intelligence Group reported the first observed fully autonomous vulnerability pipeline. A staging server hosted exploit_agent, wired to frontier and local LLMs, comparing web app code before and after patches, writing exploits, and executing them with Russian-language operator notes.
Nov 15
ChatMate and Remote Prompt Execution: Copilot Sandbox Escape ExplainedResearchers at Rubrik Zero Labs demonstrated ChatMate, the first public Remote Prompt Execution chain against Microsoft 365 Copilot. A malicious document can trigger code in Copilot's interpreter, escape Azure Dynamic Sessions, and hand an attacker a bidirectional prompt channel that runs with the victim's identity. Microsoft fixed the host escape as CVE-2026-32193 before disclosure.
Nov 15
Uber Fireworks AI Breach: Driver Licenses Stolen via Vendor KeyUber told TechCrunch on November 13, 2026 that hackers used compromised credentials at inference vendor Fireworks AI to access Uber-held driver data, including thousands of U.S. driver license records named in state breach filings.
Nov 14
Amazon MadPot Decoys Caught ARTEX AI Hacking Agents Handing Over Their Own MemoryAmazon Threat Intelligence says its MadPot decoy sensors posed as LLM inference APIs and received requests from ARTEX agents that carried their full campaign history. The records describe a self-funding AI key theft operation and failed deepfake identity checks. explainx.ai separates what Amazon saw from what the agents merely claimed.
Nov 14
DOC-foundry: Adversarial Agents That Prove Sigma Rules Before You Ship ThemDOC-foundry is a new open-source project on Hacker News that authors Sigma detection rules with LLM agents, stress-tests them with a fixed Red adversary, and ships a proof bundle with FP/FN counts and a reproducibility hash. Here is how the two pipelines differ and how to run them offline.
Nov 13
Cloudflare Modern Security Model: Six Layers Against ARTEX-Style Agentic PentestsOn November 13, 2026 Cloudflare published telemetry from the world
Nov 13
FBI ARTEX FLASH: What Is Claimed vs Verified About the Agentic AI WarningThe FBI FLASH-20261113-001 bulletin on ic3.gov warns that criminals have used the ARTEX agentic penetration-testing framework against financial, government, and technology targets since at least August 2026. This post maps what the PDF states, what CERT-EU and Microsoft already proved, and what security teams should do this week.
Nov 13
FBI FaceStealer: North Korea's $400M Deepfake Crypto Scam ExplainedOn November 12, 2026 the Justice Department and FBI detailed FaceStealer, a North Korean social-engineering campaign that stole more than $400 million in cryptocurrency using hijacked accounts, fake video meetings, stolen webcam footage, and growing use of AI deepfakes. Here is how the scam works and what builders should change.
Nov 12
Fake OpenAI, SpaceX and Anthropic Event Invites Are Stealing Microsoft and Google LoginsPush Security reported on November 5, 2026 that attackers are sending fake event invitations themed on OpenAI, SpaceX and Anthropic. Employees who click "Continue with Microsoft" or "Continue with Google" hand over tokens, passwords and session cookies, even with MFA turned on. Here is the flow, the two phishing-as-a-service tools behind it, and the controls that work.
Nov 11
AI Voices in 911 Swatting: APCO Survey and Why Dispatchers StruggleA November 2026 NPR report ties a California school bomb hoax to an AI-generated caller voice. APCO International says about 14% of surveyed 911 professionals already faced AI-enabled swatting, as NENA trains dispatchers on new detection limits.
Nov 10
NEAR AI Pauses Cloud Inference After Internal Breach, Then Restores ServiceOn November 9, 2026 NEAR AI detected unauthorized access to internal systems and paused NEAR AI Cloud inference while investigating. The company said customer workloads run in hardware-isolated TEEs and that NEAR Protocol was unaffected. On November 10 it announced the cloud was back online after fixing vulnerabilities and rotating credentials, with a full incident report still pending.
Nov 10
Notion AI Chat Transcripts Exposed in Fireworks Breach: What Notion SaysOn November 10, 2026 a Hacker News Tell HN quoted Notion's customer email about the Fireworks AI breach. Notion says under 1% of workspaces may be affected, Insights and open-weight Fireworks routing are disabled, and users should treat leaked chats as exposed.
Nov 10
Search Without AI Overviews: What You Can Actually Turn OffNov 9
Fireworks AI October 2026 Breach: Billing Data, Model Weights and LLMjacking UpdateOn November 9, 2026 Fireworks AI expanded its October incident bulletin. The company now says billing addresses and invoice history were accessed, some incomplete model weights and datasets were taken, and a second chain on October 23 went through a cloud provider into R&D machines. Inference and zero-retention paths were not impacted per Fireworks.
Nov 9
Sleep Wearables, EEG Data and Brain Foundation Models Without Clear ConsentA Science investigation with the Pulitzer Center reports that consumer sleep and meditation headbands collected tens of millions of hours of EEG data, while several firms trained brain foundation models without clearly telling users in privacy policies.
Nov 8
Human Slop: AI Is Exposing Decades of Neglected Software BugsA Business Insider essay reframes Black October: AI is not inventing most of the crisis, it is scanning old human-written software faster. We connect that argument to PHP2Shell, open-source maintainer strain, and defender initiatives readers can act on.
Nov 8
Microsoft Links Cambodia's Storm-3358 to ARTEX Attacks on Financial FirmsMicrosoft Threat Intelligence published a November 6, 2026 report naming Storm-3358, a Cambodia-based financially motivated cluster, as the latest group to weaponize ARTEX against banks and other financial firms. explainx.ai lays out what Microsoft confirms, how it differs from Storm-0940, and what security teams can do about machine-speed scans.
Nov 7
Alex Stamos on Decoder: AI Cybersecurity Dark Age and What Defenders Do NowOn November 7, 2026, Alex Stamos told Decoder that cheap open-weight models can automate most of the cyber kill chain, that breach chains like Fireworks-to-Cursor are the new normal, and that defenders face a dark age until SOC work gets cheaper than attacker tokens.
Nov 7
ChatGPT Silently Switched Connectors in a Follow-Up Tool Call, and the Bounty Wanted an IDA developer reported on the OpenAI community forum that ChatGPT routed a confirmation follow-up through a second connector pointing at the same MCP server, without telling the user. The server rejected the call, but the report raises a real design question, and the bounty process asked for an ID card.
Nov 7
Let's Encrypt AI: Confidential Computing Goal for Private Chats by 2027A new site at letsencrypt.ai publishes a position paper on confidential computing for AI inference — not HTTPS certificates from ISRG. It names live services, NVIDIA attestation, and a 2027 adoption goal.
Nov 6
Censys: 1,465 Internet-Exposed ARTEX Consoles — Many With Open Directories and Dual-Use ToolsSilas Cutler at Censys published detailed telemetry on November 5, 2026: 1,465 publicly reachable ARTEX consoles, roughly 815 that do not look like mass-cloned hosting templates, 155 that co-host other detected threats, and dozens of short-lived open directories leaking post-exploitation kits — next to Microsoft’s Storm-0940 attribution of ARTEX use in the ConnectWise supply chain.
Nov 6
Chainalysis Breach: Password Hashes and Live Tokens for All CustomersChainalysis first called its October 30 incident limited to profile data. By November 4 it said password hashes and live tokens for every user were exposed. Here is the timeline, what is confirmed, and what customers should do.
Nov 6
Check Point Reads ARTEX Operator Logs: 1,608 Live Servers, Mostly Autonomous AttacksCheck Point Research published the first operator-side look at ARTEX, the open-source AI penetration-testing platform tied to bank and backup attacks. The logs show cheap, almost fully autonomous campaigns, and results that were far weaker than the volume suggests.
Nov 6
DOJ Charges 3 Men Over $129M AI Impersonation Scheme: CEOs and Fake LawyersUS prosecutors say three men used AI to mimic executives and attorneys and talked finance staff into wiring about $129 million. Here is what the indictment alleges, what is still unproven, and the controls that stop this attack.
Nov 6
Vals AI CyberOffense: Open-Weight DeepSeek Nearly Matches Restricted GPT-6 Sol DaybreakVals AI tested seven models as autonomous attackers on two simulated corporate networks. The restricted GPT-6 Sol Daybreak led, and a downloadable DeepSeek model came close for under $3 a run. Here are the scores, the method, and what the result does not prove.
Nov 5
Cyber AI Models Compared: GPT-6 Cyber, MAI-Cyber-1, Fugu-Cyber, Gemini CyberNov 5
GPT-6 Cyber vs AgentBeam: Finding Bugs vs Controlling AgentsNov 5
Railway Outage: Forged Source IPs Got Its Addresses BlockedNov 4
Cloudflare Zero Wire: A Free Briefing Hub for Critical CVEs, With a Prompt You Can Hand to an AI AgentNov 4
Pangram Image 1.0 and Image Flash 1.0: What the AI Image and Video Detector Claims, and What Is VerifiedNov 3
Binance Says Fake Job Interviews and Internal AI Tools Led to the $6.3 Billion Hack: Claimed vs VerifiedNov 3
BLA (Blind LLM Agents): The Trending Shield Against AI Reverse-Engineering Agents, Claimed vs VerifiedNov 3
GitBook Says an Attacker Using AI Agents Took Infrastructure Credentials: What to Rotate NowNov 3
Google GTIG on LLMjacking: Stolen Cloud AI Access Now Costs Victims Up to $1 MillionNov 3
NCSC Warns AI-Powered Ransomware Will Hit UK Critical Infrastructure by 2028Nov 1
AiSOC: An MIT-Licensed AI Security Operations Center You Can Self-HostNov 1
November 2026 Cyberattacks: The Full List of Incidents, Counted by Type
October 2026
Oct 31
Postgres 19 Is Late Because LLMs Found Too Many Bugs, Says Committer Robert HaasOct 31
Starbucks Drive-Thru AI: What Niccol Said vs the "Facial Recognition" HeadlinesOct 31
VOLTA Infostealer Uses an LLM to Rank Victims by What They Are WorthOct 30
VAPOR PANDA Ran ARTEX and DeepSeek Against 200+ Organizations: What CrowdStrike ConfirmedOct 30
Epoch: An Open-Weight Model Turned Firefox Patch-Gap Fixes Into Working Exploits for About $1,000 a ReleaseOct 29
DHS Icepick Leak: 41,000 Employees Exposed, Okta Claim DisputedOct 29
Hacked AI Hardware Can Make Chatbots Give Dangerous Answers: What Eclypsium and Nvidia ShowedOct 29
Inside Proton's AI Push: What 404 Media Reports, and What Is ConfirmedOct 29
Incident Response Backlogs After Black October: What Sophos Said and What Is VerifiedOct 27
Fuzzilli vs BigSleep: Project Zero Finds AI Agents and Fuzzers Tie on V8 BugsOct 27
What Is a Zero-Day? Vulnerabilities, Exploits and Why AI Changes the ClockOct 26
Cohere Says Qwen3.8-Flash-Next Tried to Upload a File to a Beijing BucketOct 26
Grand Theft AI: Highway Thieves Are Stealing Millions in Data Center GearOct 26
Tailworm: What the Tailscale SSH Flaws Mean If You Run AI AgentsOct 25
GMO Internet Cloud Breach: What Is Confirmed and What Is Claimed About ARTEXOct 25
How to Blur Faces and Backgrounds on iPhone Duo (AI Privacy Guide)Oct 25
A $50 Backdoored Qwen2.5-7B Stole .env Credentials Through Codex CLI: What ProjectDiscovery ShowedOct 24
Brotli Vulnerability CVE-2026-111497: Why It Is Bigger Than Node.jsOct 24
How to Learn Cybersecurity from Zero in 2026: A 9-Level Roadmap for the AI EraOct 23
Anthropic Deputy CISO Warns It Will Get "A Lot Worse" for 18 Months: What Is Confirmed and What to DoOct 23
The Internet Is Under Attack: How to Protect Yourself, and Why the SaaS Era Is Drying UpOct 23
Internet Safety for Everyday People: A Simple Guide to Hacks, Scams and UpdatesOct 23
October 2026 Cyberattacks: The Full List of 106 Incidents, Counted by TypeOct 22
Archestra 1.4 "Arpa": OpenAPPA Guardrails, a Remote Agent Runtime and RBAC for Claude Code and CodexOct 22
CyberCartel Used Claude, Grok and Kimi Agents Against 400+ Latin American Financial Targets: What Is ConfirmedOct 22
eMudhra CA Compromise: The "Update Everything" Security Checklist and What a Viral Thread MissedOct 22
Microsoft Ties ConnectWise Attack to China-Linked Storm-0940 and AI Tool ARTEX: What Is ConfirmedOct 22
Railway Hack and the OpenRouter Breach Claim: Why Attackers Are Stealing AI API Keys, and What to DoOct 22
Symantec Finds "Skynet Mode" LLMjacking Agent That Hijacks Exposed AI ModelsOct 22
Orélien, Goncourt and the AI Detector: Did the Publisher "Drop" the Author? Claimed vs VerifiedOct 22
Trail of Bits Releases patching-bugs and regression-hunt After 21 Regressions in 130 AI-Assisted PatchesOct 21
eMudhra Explained Simply: Why a Secret Key Equal to "4" Shook Website Security, and What to DoOct 21
LiteLLM 1.107.0 on PyPI Is Malicious: What the Credential Stealer Takes and How to Clean UpOct 21
Node.js Security Updates, October 2026: What Was Fixed, Who Should Update, and HowOct 21
Why the NSA Is Still Telling Everyone to Turn Off Cisco Smart Install, Explained SimplyOct 21
PHP2Shell Explained Simply: The AI-Found PHP Bug Behind Today's Emergency Patches, and What to DoOct 20
ECHOTRIBBLE: The Army Cyber Command Program Building Autonomous AI Hacking AgentsOct 20
Spotify Pulled 50 Million AI Slop Tracks From Recommendations: What the Numbers MeanOct 20
Yuki Chen Joins Z.ai: 23 Microsoft Flaws Credited "Using GLM" — Claimed vs VerifiedOct 19
One Hacker, AI Agents and About $5,000: How Deere, SharkNinja and Cummins Sites Were BreachedOct 18
ARTEX Global Edition: The English Fork of the Pentest Agent Tied to the Korean Bank HacksOct 18
What Is a Data Breach? Types, Real Examples, and What to Do (AI Era Guide)Oct 16
Calif.io Jailbreaks iOS 27.0.1 on iPhone 18 Pro: What Is Claimed vs. VerifiedOct 15
Brave Leo Now Encrypts AI Chats End to End: How OHTTP, NEAR AI TEEs and GLM 5.3 Flash Keep Brave OutOct 14
AI Agents Are Not "Going Rogue": Experts Say Frontier Labs Skip Security BasicsOct 14
Apate Sends AI Agents Undercover to Get Scammed and Harvest Fraud IntelligenceOct 14
Orion-1: A Cyber Defense Model Built Because Guardrails Block DefendersOct 14
PentAGI: The Open-Source Autonomous Pentest Agent Trending on GitHubOct 13
Bill Gates Warns of "Negative Surprises" From AI Cyberattacks Within a Year: What He Said and What the Data ShowsOct 13
Raven: The AI Tool Police Use to Find the Exact House in a Single Interior PhotoOct 13
PS5 Relapse Jailbreak and Overdrive: Did AI Really Crack It? Claimed vs. VerifiedOct 13
XBOW Built a Chrome Full-Chain Exploit in 6 Days With GPT-5.6 Sol and Won $250,000Oct 12
The $442 Billion Scam Economy: What Bloomberg Claims and What the Data SupportsOct 10
FBI Arrests Ransomware Negotiation Firm Co-Founder in ShinyHunters ProbeOct 9
Flock Safety to Cut 18% of Staff as License-Plate Surveillance Backlash GrowsOct 9
Japan Cyberattacks September 2026: Times Car, Gyazo, Keio and Monogatari ExplainedOct 8
Can AI Break Cryptography? What Is Claimed vs. What Is VerifiedOct 7
Armadin's AI Attack Swarms Reportedly Found 90+ Zero-Days at Fortune 500 FirmsOct 6
Security-One 27B: An Open Decision Model for Prompt-Injection and Security TriageOct 6
South Korea Bank Hacks: What AI Actually Did, and What Builders Should LearnOct 3
Ethereum zkAPI: Pay for AI APIs in ETH or USDC Without Linking Requests to You
September 2026
Sep 30
.si Domains: What They Are, How Registration Works, and How to Avoid ScamsSep 26
The Provenance Tax: how LLM watermarking can break agent tool calls and refusalsSep 24
CLOSEDQUORUM: Cisco Talos Finds Windows Malware That Lets Four AI Models Vote on Its Next MoveSep 24
VS Code Remote-SSH Is Not a Sandbox: What the Fly.io "Bananas" Post Gets Right, and What Hacker News Argued AboutSep 23
Microsoft and Coinbase Dismantled an AI Cybercrime Platform Called EvilTokensSep 23
Z.ai Open-Sourced ZCode After Its Default Config Uploaded Code to Alibaba CloudSep 22
"Spymarks": The Case for a New Word for Hidden AI TrackingSep 21
Did Claude Factor RSA-896? What the Bare Headline Actually ClaimsSep 19
AgentCloak: Swap Your Real Data for Fakes Before Any AI Sees ItSep 19
Plugin4Shell: A Zero-Click RCE Hit Claude Code, Codex, Copilot, and Gemini CLISep 19
Top AI Agent Safety Workshops in 2026: Free and Paid, ComparedSep 18
Researchers Chained a libheif Bug and an OpenAI SSO Flaw — With ClaudeSep 17
Kalypta: The App That Blocks AI Notetakers From Your MeetingsSep 17
Pangram Launches a Gmail AI Labeler With a 1-in-10,000 False Positive RateSep 10
LG TVs Caught Recording Audio Even When "Off," Gamers Nexus FindsSep 9
Boris Cherny Benchmarks GPT-6 Astra's Prompt Injection ResistanceSep 9
Check Point Found a ChatGPT Sandbox Flaw That Leaked Gmail Across AccountsSep 6
AI Agents Can Now "Hand-Draw" Art — and Fake the Timelapse TooSep 5
Instinct + 1Password: What It Actually Means to Give an AI Agent Your VaultSep 5
numbat: Perplexity's Open-Source Observability Tool for AI AgentsSep 3
Reported Git Exploit Runs Code Before the Trust Prompt in 7 AI AgentsSep 3
RSA-260 Just Fell. We Verified It in Three Lines of Python.Sep 2
Claude Fable 5.1 System Prompt Leak: Did It Really Expose Private Memories?Sep 1
CrowdStrike Falcon IQ Deploys 50+ Agents for AI Risk AssessmentsSep 1
Is Writing the Safest Job From AI? Mollick and Demirbas Both Say the Party's Over
August 2026
Aug 29
AI Watermark Removal: What Is Right, What Is Wrong, and What Actually WorksAug 29
Smart Glasses Misuse, Venue Bans, and How to Protest in 2026Aug 27
Core Lightning's AI-Found Bugs: What Actually Happened (Not "Shutdown")Aug 26
C2PA Android Cameras Broken: Pixel Assurance Level 2 Forged AnywayAug 26
Fake Codex Installer: Google Ads ClickFix Delivers AMOS on macOSAug 24
seL4 on AArch64: Proofcraft Completes the Security Proof StackAug 21
What Is C2PA? Content Credentials, ExplainedAug 21
What Is Indirect Prompt Injection? How Web Content Hijacks AI AgentsAug 20
CISA Warns AI-Generated Exploits Are Targeting Siemens S7 PLCs Right NowAug 20
Grok Leaks Private Chats via Encrypted Prompt Injection — Guardrails Can't Read AESAug 18
How to Disable AI Features in Windows, Chrome, Zoom, and MoreAug 18
Top 10 Signs of AI-Generated Text (2026)Aug 17
The AI Credit Resale Market: Is Cheap Claude/GPT Access Safe?Aug 15
Google HEIR: A Compiler for Running AI Inference on Encrypted DataAug 14
ExploitBench: The Benchmark Measuring How Far AI Can Exploit Real CodeAug 13
A Watermark Removal Tool Just Added OpenAI and Gemini SupportAug 12
What AI Watermarking Actually Changes for DevelopersAug 12
What AI Watermarking Actually Changes for MarketersAug 12
How Does AI Text Watermarking Actually Work? A Technical ExplainerAug 11
Kimsuky Ran LLMs Offline on Its Own Servers — What That Breaks for DefendersAug 11
Stealing Reasoning Traces: The Encrypted Chain-of-Thought Flaw in Every Frontier LLM APIAug 10
How to Restrict What Claude Desktop Can Access on Your ComputerAug 10
tl;dv Data Breach: 181,874 Meetings Exposed, Live Calls JoinableAug 5
Microsoft Orchard: Open-Source Agentic Modeling Framework Explained
July 2026
Jul 29
Copilot for Word AI Worm: Document-Borne XPIA That Self-PropagatesJul 27
Token Relay Market: How Cheap Claude Access Gets SoldJul 27
GrapheneOS Duress Wipe Now the Center of a Federal ProsecutionJul 24
Cisco Antares: Open-Weight SLMs for Vulnerability LocalizationJul 23
Substack Launches AI Detection with Pangram — What It Flags and HowJul 20
AI Cyber Guardrails Block US Defenders — Kimi K3 and GLM 5.2 Fix What Codex and Fable RefusedJul 17
LLM Text Detection with Classical ML — TF-IDF + SVM That Still Works (2026)Jul 9
EU Driver-Facing Camera Law (2026): ADDW Privacy Guide for Drivers and BuildersJul 8
AI Found 7 Bugs in Cloudflare CIRCL: What zkSecurity's zkao Audit Reveals
On July 7, 2026, zkSecurity reported that its LLM pipeline and zkao audit agent confirmed seven real vulnerabilities in Cloudflare's CIRCL cryptography library, all since fixed upstream. They included float64 precision loss in threshold RSA, a BLS rogue-key flaw and a CP-ABE access-control break found by zkao. The AI often misjudged severity, so human triage stayed the bottleneck across 200+ scanned projects and over 1000 candidate findings.
Sources: zksecurity.xyz, github.com, zksecurity.xyz
Jul 8
GitLost: GitHub Agentic Workflows Leaked Private Repos via Prompt InjectionJul 8
Meetily: Privacy-First AI Meeting Assistant With Local Whisper and Parakeet
June 2026
Jun 30
WhatsApp Usernames: Reserve Your Handle Before the 2026 RolloutJun 29
SimpleX Chat: The Only Messenger With No User Identifiers — Setup Guide 2026Jun 23
Trump's Quantum Executive Orders: A 2028 Quantum Computer, a 2030 Encryption Deadline, and What Developers Need to KnowJun 17
Flock Safety, ALPRs, and the AI Surveillance Debate: Civil Liberties, Law, and the Cameras Watching Every Car in America (2026)Jun 15
NVIDIA SkillSpector: Security Scanner for AI Agent Skills (2026)