Off the Wire
What’s New
New and updated since 19 October 2026. The latest stories filed to the Digest, newest first.
New since last issue
439 stories- Unpatched Tailscale SSH Flaws Could Let a Hacked Device Reach Root on Other Servers25 Oct
- PHP2Shell Exploitation Hits Government Entity Amid Global Scanning and Signs of AI-Assisted Hacking25 Oct
- OpenAI AI Agents Demonstrate VM Escapes and Secret Cross-Sandbox Communication25 Oct
- BlackX UEFI bootkit exploits Baton Drop flaw to bypass Secure Boot24 Oct
- Brotli Decoder Flaw Behind Node.js RCE Also Affects Bun, Python and Other Widely Used Software24 Oct
- Google Says Gmail and Photos Data Shared With AI Search Could Be Used for Targeted Ads24 Oct
- Trust Is Not What It Seems — And Fraudsters Are Counting on It23 Oct
- Abandoned Popcorn Time app exposed millions of devices to RCE attacks23 Oct
- Microsoft Teams to get AI-powered, context-aware phishing detection23 Oct
- Exposed Infrastructure Reveals How Ransomware Affiliates Used AI Agents to Execute Intrusions23 Oct
- FCC bans two devices with surveillance chips linked to Huawei23 Oct
- AI Tools Are Generating Massive Attacks for Little Cost23 Oct
- The Good, the Bad and the Ugly in Cybersecurity - Week 43 (2026)23 Oct
- Microsoft: Windows Server 2016 reaches end of support in January23 Oct
- Quest Diagnostics confirms data breach exposed patient records23 Oct
- Signed, Sealed, and Relayed, NTLM-style: Bypassing LDAP Signing Through Reauthentication23 Oct
- You gave your AI agent access, but who is giving it orders?23 Oct
- Hackers Breach Enterprise Firewalls and VPNs, Exposing Over 70,000 Credentials23 Oct
- A Protected Browser Doesn't Mean Your Enterprise Is Remediated23 Oct
- UK proposes 14-year prison terms for undersea cable sabotage23 Oct
- HPE patches eight max severity RCE and auth bypass AOS flaws23 Oct
- AI Agents as Insider Threats: Inside Microsoft's Bet on a Unified Security Stack23 Oct
- Russia-Linked Hackers Used New Mesh Network Backdoor in Destructive Ukraine Attack23 Oct
- Compromised TLS Certificates23 Oct
- [NEU] [hoch] Oracle Java SE and GraalVM, und OpenJDK: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen23 Oct
- [NEU] [hoch] TianoCore EDK2 PXE Treiber: Schwachstelle ermöglicht Ausführen von beliebigem Programmcode und DoS23 Oct
- [NEU] [hoch] Fortinet FortiOS, FortiManager, FortiAnalyzer, FortiMail und FortiAuthenticator: Mehrere Schwachstellen ermöglichen nicht spezifizierten Angriff23 Oct
- [NEU] [mittel] MongoDB C# Treiber: Mehrere Schwachstellen23 Oct
- [NEU] [hoch] Aruba ArubaOS: Mehrere Schwachstellen23 Oct
- [NEU] [UNGEPATCHT] [mittel] expat: Schwachstelle ermöglicht Denial of Service23 Oct
- [NEU] [mittel] Red Hat Enterprise Linux ca-certificates: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen23 Oct
- [UPDATE] [hoch] Apple iOS, iPadOS, macOS Tahoe, macOS Golden Gate, macOS Sequoia und Safari: Mehrere Schwachstellen23 Oct
- [UPDATE] [hoch] WebKitGTK: Schwachstelle ermöglicht Codeausführung23 Oct
- [UPDATE] [hoch] Android Patchday November 202423 Oct
- [UPDATE] [mittel] Oracle Java SE: Mehrere Schwachstellen23 Oct
- Critical Flaw in Python’s Popular Pillow Image Library Allows Code Execution via a Crafted Image23 Oct
- Researchers Demonstrate Node2Shell RCE Exploit on Node.js Build Using Crafted Brotli Data23 Oct
- Critical x402 SDK Flaw Could Let Malicious APIs Drain AI Agents’ Token Balances23 Oct
- JANUS: Agentic Cross-Language Fuzzing Harness Generation for Android Native Libraries23 Oct
- Automated screening of privacy risks in medical AI23 Oct
- Not Where You Think: Protecting Location Privacy Against Reasoning-Based Geolocation23 Oct
- Social-HUMINT: An LLM-Augmented Unified Framework for Measuring Social Media Exposure23 Oct
- AgentPhishing: A Benchmark of LLM Email Agents Under Adaptive Attack23 Oct
- AfterDeleteBench: Auditing Post-Deletion Information Persistence in LLM Agents23 Oct
- Practical Auditing of Differentially Private Graph Release Mechanisms: Methodology, Toolkit, and Case Study23 Oct
- Weaponizing the Exit: Input Privacy Leakage via Exit Routing Manipulation in Multi-Exit Networks23 Oct
- SecLog-Bench: A Scalable Benchmark for Cybersecurity Log Reasoning with Scaffolded Evidence23 Oct
- ReDence: A Hybrid Agentic and Decision-Model Framework for Rapid, Evidence-Grounded Detection of Modern Social Engineering Scams23 Oct
- MPIDetector: Detecting Multi-modal Prompt Injection via LLM-based Reasoning23 Oct
- MIG-COW: Transferable Adversarial Attacks on Deepfake Detectors via Gradient Decomposition23 Oct
- Amplifying Latency and Cost of LLM Inference via Rowhammer Attack23 Oct
- AIRO: A Streaming-Based Machine Learning Framework for Threat Detection and Security in Industrial Internet of Things23 Oct
- Your Token Is Gone: A Token-Consumption Countermeasure for Disrupting LLM-Based Scam Systems23 Oct
- Model-intrinsic, Weight-level, and Training-free Spectral Fingerprinting for Large Language Models23 Oct
- HiMark: Hierarchical and Controllable Multi-bit Watermarking for Large Language Models23 Oct
- Bridging the Closed-World Gap: Interpretable Anomaly Detection via a Hybrid TSK-FCM Ensemble on the BETH Dataset23 Oct
- Beyond End-to-End Recovery: Towards a Better Understanding of LLM Steganography via Rolewise Decomposition23 Oct
- Membership Inference Attack on Private Evolution23 Oct
- Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems23 Oct
- HAWC-Leak: Heterogeneous Architecture Workload Classification utilizing Side-Channel Leakage23 Oct
- Rank Ceiling for Twiddle-Perturbation Faults on the Forward NTT23 Oct
- Data-Void Exploits in Web-Searching LLMs23 Oct
- Can Loan Processors and Underwriters Spot AI-Edited Statements? An Empirical Study23 Oct
- Structural cryptanalysis of classic McEliece: Assessment of claims, costs and evidence23 Oct
- AgentTriage: Replayable Security Diagnosis Across Heterogeneous Agent Runtimes23 Oct
- Scaling Laws for Safety Probes23 Oct
- FVQKD: A Formal Verification of Quantum Key Distribution Security in Rocq23 Oct
- WildFakeToolBench: Benchmarking AI-Generated Image Detectors with Web-Sourced Image Processing23 Oct
- Who Patches the Lightning Network? Evidence from Release Lines in Gossip, 2023-202623 Oct
- Mirage: Targeted Jailbreaks Against Post-Generation Safety Checkers in Multimodal Diffusion Models23 Oct
- PhishGen: Understanding LLM Capabilities in Automated Phishing Website Generation23 Oct
- Measuring the Permission Gate: A Stress-Test Evaluation of Claude Code's Auto Mode23 Oct
- Beyond Attack Success Rate: Cost-Aware Metric for Evaluating Iterative LLM Jailbreaks23 Oct
- ARMOR: Adversary Cost Amplification via a Schema-Preserving Password-Hardening Retrofit for Resource-Constrained Systems23 Oct
- SecDT: A Profile-Based Security Layer for TRDP Communications23 Oct
- Metamorphism: A mathematical challenge for antivirus technology23 Oct
- Does More Compute Make Every Adversarial Attack Transferable?23 Oct
- An Orchestrated Sandbox for Side-Channel Analysis of Cyber-Physical Systems Malware23 Oct
- No Edges, No Verdict: A Large-Scale Empirical Study of Declared Dependency Graphs in 78K SBOMs in the Wild23 Oct
- How to Backdoor Image Knowledge Distillation23 Oct
- Railway warns devs to rotate secrets after potential breach22 Oct
- Capcom confirms data breach affecting 11.7 million accounts22 Oct
- Les supermarchés de l'UE paient d'un firmware de réfrigérateur non mis à jour22 Oct
- LiteLLM release pipeline hacked to distribute info-stealer malware22 Oct
- Indian tech company eMudhra announces withdrawal from certificate business after security incident22 Oct
- eMudhra to Exit Global SSL/TLS Business After 8 Intermediate CA Keys Found Easily Guessable22 Oct
- I had plans for that second22 Oct
- Arista patches new VeloCloud Orchestrator zero-day exploited in attacks22 Oct
- Two CVSS 9.0 Linux KSMBD RCE Flaws Disclosed on Public Mailing Lists Before Kernel Fixes22 Oct
- The Current Landscape of OAuth Attacks22 Oct
- Microsoft blames ConnectWise attacks on Chinese state hacker using ARTEX AI agents22 Oct
- [tl;dr sec] #351 - PHP2Shell, Autopatching Skills, AWS Pathrunner22 Oct
- [UPDATE] [mittel] Red Hat OpenShift Container Platform (protobufjs, fast-uri): Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Google Cloud Platform (GKE containerd): Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Golang Go: Mehrere Schwachstellen ermöglichen nicht spezifizierten Angriff22 Oct
- [UPDATE] [hoch] PHP: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Golang Go: Mehrere Schwachstellen22 Oct
- [UPDATE] [mittel] Golang Go: Mehrere Schwachstellen22 Oct
- [UPDATE] [mittel] VMware Tanzu Spring Security: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen22 Oct
- [UPDATE] [hoch] PostgreSQL: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Apache HTTP Server: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] PHP: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] PHP: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] PHP: Mehrere Schwachstellen22 Oct
- [NEU] [mittel] Node.js: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Internet Systems Consortium BIND: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] MongoDB Compass: Schwachstelle ermöglicht Codeausführung22 Oct
- [NEU] [hoch] Apache Traffic Server: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Red Hat Ansible Automation Platform: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Vaadin: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Cisco Produkte (Emergency Responder, Prime Collaboration, Unified Communications Manager und Unity Connection): Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Cisco Expressway: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Cisco Produkte (IOS XE, SFTD) und Snort: Mehrere Schwachstellen22 Oct
- [NEU] [kritisch] GStreamer: Schwachstelle ermöglicht Codeausführung22 Oct
- [NEU] [hoch] Splunk Splunk Enterprise: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Kyverno: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] JupyterHub: Mehrere Schwachstellen22 Oct
- [NEU] [hoch] Splunk SOAR: Mehrere Schwachstellen22 Oct
- [NEU] [mittel] WSO2 API Manager: Mehrere Schwachstellen22 Oct
- As AI Risks Grow, Insurers Split on Covering the Damage22 Oct
- Exposed Refrigeration Systems Raise Global Cybersecurity Concerns22 Oct
- New Serious PHP Vulnerabilities Released and Patched22 Oct
- Amazon says PHP2Shell exploitation began 4 days before patches22 Oct
- ARTEX in the Wild: Storm-0940 Expands Targeting of IT and Software Vendors22 Oct
- [UPDATE] [hoch] Eclipse Jetty: Schwachstelle ermöglicht Darstellen falscher Informationen und Offenlegung22 Oct
- [UPDATE] [hoch] PHP: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Bouncy Castle: Mehrere Schwachstellen22 Oct
- [UPDATE] [mittel] Red Hat OpenShift Container Platform (fast-uri,OpenTelemetry-Go) : Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Red Hat Satellite (foreman, python-pillow, go): Mehrere Schwachstellen22 Oct
- GitLab warns of new max severity code execution vulnerability22 Oct
- Node.js Brotli Decoder Bug Could Allow Remote Code Execution in All Supported Versions22 Oct
- New QPACK Flaws Can Crash HTTP/3 Servers, Trigger Up to 4,000-Fold Memory Amplification22 Oct
- [UPDATE] [mittel] Wireshark: Mehrere Schwachstellen ermöglichen Denial of Service22 Oct
- [UPDATE] [hoch] Drupal Erweiterungen: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] NoMachine: Mehrere Schwachstellen22 Oct
- [UPDATE] [hoch] Drupal Erweiterungen: Mehrere Schwachstellen22 Oct
- Sign Here: Attributing a Source for a More Trustworthy Software Supply Chain22 Oct
- MASAR: Multi-stage Attack Surface Analysis and Reasoning Framework for Telecom Networks22 Oct
- Characterizing and Defending Utility-Hijacking in Instruction-Tuning Data Selection22 Oct
- Compaction Laundering: Provenance Destruction in Context-Compacting Coding Agents22 Oct
- Black-Box Red-Teaming of Multi-Agent Systems via Reinforcement Learning22 Oct
- How Much Does the Harness Matter, and for Whom? A Multi-Model Study of Context in LLM-Driven Vulnerability Discovery22 Oct
- Where AI Vulnerability Discovery Harnesses Go Quiet: Structural False Negatives and Hidden Vulnerabilities22 Oct
- CertMoE: Robustness Certification for Mixture-of-Experts LLMs via Routing-Aware Smoothing22 Oct
- VMSpec: A Decompiler-based Approach against Virtual Machine Obfuscation22 Oct
- Beyond Policy Engines: Binding-Complete Execution Admission22 Oct
- SUDP: Secret-Use Delegation Protocol for Agentic Systems22 Oct
- VDDP: Verifiable Distributed Differential Privacy under the Client-Server-Verifier Setup22 Oct
- Do Flow-Generation Tools Matter? Unsupervised Characterization of MQTT Traffic Across Public Datasets22 Oct
- PQ-Vault: Provably Quantum-Safe Asset Custody for Unmodified Ethereum22 Oct
- Combining Long-Term Memory and LLMs Comes at a High Cost: From Prompt Injection to Persistent Compromise22 Oct
- Using AWS WAF to apply mitigations for PHP2Shell RCE vulnerabilities (CVE-2026-109618 and CVE-2026-109619)22 Oct
- Google Chrome adds Gemini-powered scam detection, warns users of ClickFix attacks21 Oct
- Japan's Onamae Desktop Cloud hacked, up to 27,000 servers affected21 Oct
- TiDB Breach Exposes Flawed Cloud Security Assumptions21 Oct
- Lazarus hackers use post-quantum cryptography for malware delivery21 Oct
- Marvell, Microsoft & Oracle Work to Standardize Data Center Hardware Security21 Oct
- Cloud Security Alliance Survey Finds Data Loss Prevention Often Detects Data Loss Only After It Happens21 Oct
- CISA: Hackers chain Plex flaws in Plex2Shell attacks to breach servers21 Oct
- The Limits of Fingerprint Blocklists: A Reproducible Java Certificate Bypass21 Oct
- PHP says AI review of critical bugs first sparked alarm in August, but signs surfaced in 202421 Oct
- eMudhra CA Incident: Sonatype's Assessment and Guidance21 Oct
- Threat Brief: Critical Vulnerabilities in PHP: CVE-2026-109618, CVE-2026-10961921 Oct
- OpenAI agents disabled safety monitoring, one lied about it21 Oct
- Widespread PHP Flaw May Cause Many Site Outages Today21 Oct
- Two critical RCEs in PHP core21 Oct
- Your AI agent fixed the bug, but what else did it change?21 Oct
- Hundreds of websites taken offline as concern grows over two PHP vulnerabilities21 Oct
- “Next Exploit Opportunities”: How a Threat Actor Used AI to Help Plan and Execute Multiple Types of Cybercrime Activities21 Oct
- ClickFix has never been bigger, and AI tools are its most convincing lure yet21 Oct
- Decades-Old Windows Utility Powers Modern ClickFix Attack21 Oct
- Dutch National Cyber Security Center is Offline because it uses PHP21 Oct
- China-Aligned Bronze Sandcastle Abuses Sandboxie DLL Sideloading to Deploy Backdoors21 Oct
- Generative AI Is Erasing Behavioral Tells, Weakening Behavior-Only Detection, Proofpoint Reports21 Oct
- [NEU] [hoch] Oracle Communications: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Systems: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle E-Business Suite: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Enterprise Manager: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Financial Services Applications: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Retail Applications: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Fusion Middleware: Mehrere Schwachstellen21 Oct
- [NEU] [mittel] Oracle Java SE: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle MySQL: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] vllm: Mehrere Schwachstellen21 Oct
- [NEU] [mittel] PowerDNS Recursor: Mehrere Schwachstellen21 Oct
- [NEU] [mittel] Unbound: Mehrere Schwachstellen21 Oct
- [NEU] [mittel] Checkmk: Schwachstelle ermöglicht Privilegieneskalation21 Oct
- [NEU] [mittel] FreeBSD Project (caroot): Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen21 Oct
- [NEU] [hoch] FreeRDP: Mehrere Schwachstellen21 Oct
- [NEU] [kritisch] DrayTek Vigor: Mehrere Schwachstellen ermöglichen Codeausführung21 Oct
- [NEU] [hoch] Mozilla Produkte (Thunderbird, Firefox ESR und Firefox): Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Database Server: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Apache Tomcat: Mehrere Schwachstellen21 Oct
- [NEU] [hoch] Oracle Virtualization: Mehrere Schwachstellen21 Oct
- [NEU] [mittel] Red Hat Advanced Cluster Security for Kubernetes (RHACS): Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen21 Oct
- [NEU] [hoch] Oracle REST Data Services: Mehrere Schwachstellen21 Oct
- [UPDATE] [mittel] Grafana: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen und Offenlegung von Informationen21 Oct
- [UPDATE] [mittel] CPython: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen21 Oct
- [UPDATE] [hoch] WebKitGTK: Mehrere Schwachstellen21 Oct
- [UPDATE] [mittel] Grafana: Mehrere Schwachstellen21 Oct
- [UPDATE] [hoch] Red Hat Enterprise Linux (python-pip): Schwachstelle ermöglicht Codeausführung21 Oct
- [UPDATE] [hoch] Argo CD: Schwachstelle ermöglicht Offenlegung von Informationen21 Oct
- [UPDATE] [hoch] Red Hat Enterprise Linux (git-lfs, opentelemetry-collector): Mehrere Schwachstellen21 Oct
- The Harness Is the Hard Part of Agentic Pentesting21 Oct
- Attackers Hijacked Over 300 Tor Exit Relays to Intercept SSH Connections21 Oct
- Microsoft says October updates didn’t actually fix VPN issues21 Oct
- Parasitizing trust: How legitimate infrastructure becomes a delivery pipeline for attackers21 Oct
- [NEU] [UNGEPATCHT] [hoch] expat: Schwachstelle ermöglicht Denial of Service21 Oct
- Hackers Hid Vasilek Backdoor in VMware Tools During 2-Year Russian Healthcare Intrusion21 Oct
- PHP session.upload_progress heap UAF (CVE-2026-109618 / -109619)21 Oct
- Model Exposure on One Device Endangers Others: Rethinking Edge Neural Network Security across Devices21 Oct
- KNOX: Hardware-assisted Regulatory Sandbox for Generative Models21 Oct
- Penumbra: Sample-Efficient Adversarial Search for Regulatory Obligations21 Oct
- SPECTER: Covert Data Exfiltration from Air-Gapped Systems via USB Power Modulation21 Oct
- Adversarial Benchmarks and BenchDoor Attacks: Exploiting Benchmark Incentives for Backdoor Attacks21 Oct
- SRE: Accelerating Zero-Knowledge Proofs via Scalar Redundancy Elimination in Multi-Scalar Multiplication21 Oct
- After the Attacker Leaves: Benchmarking Carryover Attacks on LLM Agents21 Oct
- Interflow: Breaking the Offline-Online Barrier for Private AI at the Edge21 Oct
- CAESAR: Mixed Reality Search-and-Rescue Platform for Cognitive Attack Evaluation21 Oct
- Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages21 Oct
- A Hybrid Information Flow Analysis Framework for Android Hybrid Apps21 Oct
- You Are the Trigger: Speaker-Activated Backdoors21 Oct
- A Digital Twin to Improve Cyber-Social Impact21 Oct
- Privacy Accounting for Empirical Denoisers21 Oct
- The Show-Up Attack: When Subagents Win Over Their Supervisor21 Oct
- Act on Your Behalf, Beyond Your Authorization: Benchmarking LLM Agent Authorization Alignment21 Oct
- NetGuise: Adversarial Traffic Generation for Early-Stage Classifier Evasion21 Oct
- Privacy-Preserving Robustness-Guided Testing21 Oct
- When Models Know but Can't Linearly Say: Relational Encoding Limits Linear Vulnerability Classification in Code LLMs21 Oct
- Raise or Check? Predicting the Necessity of Additional Fuzzing for Continuous Fuzzing21 Oct
- CyberHull: Deterministic Malicious URL Isolation via Convex Hull21 Oct
- HydraPoT: A Configurable Multi-Agent Framework for Cost-Aware LLM-Assisted Honeypots21 Oct
- Make the Adversary Spend More Energy: Generative Deception against Language-Model Attackers in Critical Infrastructure21 Oct
- Testing TLS 1.3 PQC on constrained devices and microcontrollers21 Oct
- Contextual Candour: Role-Aware Graduated Disclosure for LLMs over Mixed-Sensitivity Contexts21 Oct
- On the Vulnerabilities and Data-Free Removal of Watermarks in Batch Normalization Layers21 Oct
- ClawMMU: Protection-Coherent Virtual Memory for Stateful LLM Agents21 Oct
- Disclosure Begets Disclosure: A Self-Reinforcing Privacy Loop in LLM Agents with Persistent Memory21 Oct
- Context-Aware Prompt Injection Detection via Small Language Models in RAG Pipelines: nokast-secureRAG21 Oct
- MAS-Guard: Unified Guardrails from Individual Agents to Multi-Agent LLM Systems21 Oct
- A Unified Attack Framework against Statistical Generative Watermarks for Large Language Models21 Oct
- Execution-Time Validation of Context-Carrying Action Intents in Industrial Cyber-Physical Systems21 Oct
- Latency-Neutral Passive Cyber Monitoring for Grid-Edge Operational Technology Networks21 Oct
- Proof-Carrying Contracts: Commitments AI Agents Cannot Break21 Oct
- Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space Defects21 Oct
- Robust LLM Watermarking with Minimal Semantic Distortion for IP Protection21 Oct
- Anywhere, Any-Stymie: Remote Activation of Trojan Malware on LiDAR with Modulated Signals21 Oct
- Right Facts, Wrong Rule: Scope-Transfer Poisoning of Retrieval-Augmented Generation21 Oct
- VeriRouter: Towards Verifiable LLM Routing21 Oct
- Secure Agentic Coding through Counterexample-Grounded Feedback21 Oct
- Unpacking Security Scanners for GitHub Actions Workflows21 Oct
- Efficient and Adaptable Detection of Malicious LLM Prompts via Bootstrap Aggregation21 Oct
- QML-PipeGuard: Detecting Channel-Substitution Attacks on Quantum Machine Learning Pipelines in Untrusted Clouds21 Oct
- Privacy-Preserving Personalized Federated Learning (PPP-FL) for Medical Imaging21 Oct
- Beyond the Prompt: Recovering Hidden Concepts in Text-to-Image Models21 Oct
- Bridging Digital Deception and Physical Fidelity: A Spatio-Temporal Attack on Cognitive Radars21 Oct
- PiSAs: Benchmarking Contextual Integrity in Multi-User Agentic Systems21 Oct
- Local Allow, Global Deny: Why Per-Action Authorization Does Not Compose in LLM Agents21 Oct
- Not Self-Decidable: LLMs Cannot Draw the Boundary of What an Agent Verifier Can Check21 Oct
- RoboFake: Are Deepfake Detectors Ready for Humanoid Face Robots?21 Oct
- InterType: Breaking the Single-Function Barrier in Neural Binary Type Inference21 Oct
- How Intervention History Changes AI Agent Sabotage on Subsequent Tasks21 Oct
- Evidence-Grounded LLM-Assisted Android Mobile Forensic Reporting21 Oct
- Reliable Estimation of Differential Privacy under Selection Bias21 Oct
- BR-MoE: Budgeted Semantic Release for Privacy-Preserving Traffic Analytics21 Oct
- Energy Consumption of Post-Quantum Cryptography on Constrained and General-Purpose Architectures21 Oct
- Mind Your Boundaries: Boundary-Aware Model Obfuscation against Cross-node Attack21 Oct
- AutoRedClaw: Automated Red-Teaming for Claw-Style Agents21 Oct
- Exposing Vulnerability Monoculture Through Systemic Adversarial Robustness21 Oct
- AgenticVulHunter: Bringing Local Agentic Vulnerability Hunting to Your Terminal21 Oct
- Operation FlashHole: Lazarus breaks security, defies control21 Oct
- Stolen UpdraftCentral keys used to backdoor WordPress sites21 Oct
- Mozilla, Google, Apple block compromised eMudhra CA certificates21 Oct
- PHP warns admins of two critical flaws, emergency patches coming tomorrow20 Oct
- Threat Brief: Active Exploitation of PAN-OS CVE-2026-0257 (Updated October 20)20 Oct
- FBI says SafePay ransomware gang ‘among most active’ with more than 600 victims since 202420 Oct
- Security advisory: Critical remote code execution vulnerabilities, update now20 Oct
- Zeabur shutting down after second security breach in two months20 Oct
- 23andMe successor settles with California for $6 million over cybersecurity failings20 Oct
- Close Identity Gaps in Software Development Before Unleashing AI20 Oct
- EU cyber law won’t force fixes for fridge controllers hit in French attack20 Oct
- Hackers Abuse AI Pen-Test Tool Cairn to Automate Massive Attack Spree20 Oct
- DHS bans DeepSeek products across US federal civilian agencies20 Oct
- WhatsApp Hid Instructions in Help Pages Telling Its AI Chatbot Not to Recommend Privacy Regulators20 Oct
- [NEU] [mittel] TYPO3 Extensions: Mehrere Schwachstellen20 Oct
- What Security Teams Can Do To Tame Wild Code20 Oct
- Go Issues Emergency Update Blocking Eight CA Keys Over Undisclosed Security Concerns20 Oct
- OVHcloud CLI releases backdoored on GitHub to steal developer secrets20 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen20 Oct
- [NEU] [mittel] Apache Airflow: Mehrere Schwachstellen ermöglichen Umgehen von Sicherheitsvorkehrungen20 Oct
- [NEU] [hoch] Snipe-IT: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] vllm: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] Froxlor: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] Golang Go: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen20 Oct
- [NEU] [hoch] Roundcube Webmail: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] Laravel: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] Langflow: Mehrere Schwachstellen20 Oct
- [NEU] [mittel] FreeRDP: Mehrere Schwachstellen20 Oct
- [UPDATE] [kritisch] WSO2 API Manager: Mehrere Schwachstellen20 Oct
- [UPDATE] [hoch] Python: Mehrere Schwachstellen20 Oct
- Russian APT28 Breaches NATO Naval Shipbuilder in NATO Country, Deploys New Reflective Loader20 Oct
- [UPDATE] [mittel] CPython: Mehrere Schwachstellen20 Oct
- [NEU] [mittel] MISP und MISP Modules: Mehrere Schwachstellen20 Oct
- [NEU] [mittel] Checkmk: Mehrere Schwachstellen20 Oct
- [UPDATE] [hoch] Apache Camel: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] Camunda: Mehrere Schwachstellen20 Oct
- [NEU] [hoch] IBM Security Verify Access: Schwachstelle ermöglicht Privilegieneskalation20 Oct
- [NEU] [hoch] Red Hat Enterprise Linux (python-tornado, resteasy-core, bubblewrap, pgvector): Mehrere Schwachstellen20 Oct
- Possible new Cellebrite iPhone Radio Zero-Day20 Oct
- ClickFix Attackers Exploit Siemens Software Center Flaw to Deploy Remote Access Trojan20 Oct
- The journey of a paper unicorn: adventures inside a modern warehouse20 Oct
- Nikkei discloses third data breach in a month, 1.8M records exposed19 Oct
- WordPress SEO plugin Rank Math compromised to backdoor websites19 Oct
- Researchers Infiltrate Microsoft Tech Support Scam Operation19 Oct
- 'AdFection' Turns Ad Impressions Into Zero-Click Spyware Delivery19 Oct
- Judge denies WhatsApp’s bid to hold spyware maker in contempt for allegedly targeting users19 Oct
- Bullhorn takes staffing platform offline to contain cyberattack19 Oct
- Two more Serbian opposition figures targeted with spyware, researchers say19 Oct
- OpenCode hack exposes API keys amid AI distillation attacks19 Oct
- Ahold Delhaize confirms 'malicious actions' behind cyber outages affecting grocery brands19 Oct
- Supply Chain Zero Days: Closing the Gap Between Publication and Detection19 Oct
- Inside the Underground Market for Know Your Client Services19 Oct
- F5 releases out-of-band patches for critical NGINX vulnerabilities19 Oct
- Feds Suspect Ransom Negotiator Ran ShinyHunters Extortion Site19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Python "virtualenv": Schwachstelle ermöglicht Codeausführung19 Oct
- [UPDATE] [hoch] Red Hat Ansible Automation Platform: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Red Hat Enterprise Linux: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Red Hat OpenShift: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Python: Schwachstelle ermöglicht Manipulation19 Oct
- [NEU] [mittel] Red Hat Quay und Enterprise Linux (pypdf, urllib3, URI): Mehrere Schwachstellen19 Oct
- [NEU] [mittel] Gitea: Mehrere Schwachstellen19 Oct
- [NEU] [mittel] Linux Kernel: Mehrere Schwachstellen19 Oct
- [NEU] [hoch] MariaDB: Mehrere Schwachstellen19 Oct
- [NEU] [mittel] OpenCTI: Schwachstelle ermöglicht Umgehen von Sicherheitsvorkehrungen19 Oct
- [UPDATE] [hoch] Xen: Mehrere Schwachstellen19 Oct
- [NEU] [hoch] DBeaver: Mehrere Schwachstellen19 Oct
- [NEU] [hoch] Apache Camel: Schwachstelle ermöglicht Codeausführung19 Oct
- GhostAction Attackers Plant Credential Stealer Disguised as Font in 94 GitHub Repositories19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen19 Oct
- Bring Your Own Vulnerable Driver: What Detection Teams Should Actually Hunt19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen ermöglichen Denial of Service19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Red Hat Enterprise Linux (corosync, libevent, libsoup): Mehrere Schwachstellen19 Oct
- From Email-to-Case to the Boardroom: How Inherited Agents Break Governance Built for Software19 Oct
- Training AIs to Solve CAPTCHAs19 Oct
- [UPDATE] [mittel] Ansible: Schwachstelle ermöglicht Codeausführung19 Oct
- [UPDATE] [hoch] OpenVPN: Mehrere Schwachstellen19 Oct
- [UPDATE] [mittel] Podman: Schwachstelle ermöglicht Manipulation von Dateien19 Oct
- [UPDATE] [hoch] WebKitGTK: Mehrere Schwachstellen19 Oct
- [UPDATE] [mittel] MongoDB Server, Go Driver, Ruby Driver: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Golang Go: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] HP HPLIP: Mehrere Schwachstellen19 Oct
- [UPDATE] [mittel] Internet Systems Consortium BIND: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Unbound: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Google Chrome / Microsoft Edge: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] Google Chrome / Microsoft Edge: Mehrere Schwachstellen ermöglichen nicht spezifizierten Angriff19 Oct
- [UPDATE] [hoch] Linux Kernel: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] strongSwan: Mehrere Schwachstellen19 Oct
- [UPDATE] [hoch] PostgreSQL: Mehrere Schwachstellen19 Oct
- [UPDATE] [mittel] OpenVPN: Schwachstelle ermöglicht DoS und die Offenlegung von Informationen19 Oct
- Fake Web3 Accelerator Uses npm Packages to Spread Malware and Hijack AI Agent Traffic19 Oct
- Microsoft: October Windows Server updates break Exchange sign-in, mail flow19 Oct
- Unpatched Broadcom Wi-Fi Flaw Lets Attackers in Range Run Code as Root on Routers19 Oct
- From Billion-Scale to Million-Scale Language Models: Operational Capabilities for Embedded Post-Exploitation19 Oct
- Monitor-Validity Failures: LLM Safety Evidence Paths at Evolving Framework Boundaries19 Oct
- A Deeper Dive into the Irreversibility of PolyProtect: Making Protected Face Templates Harder to Invert19 Oct
- CT0: Cybersecurity Transformer19 Oct
- Broad and Efficient PII Detection with Unified Labels19 Oct
- RedDarwin: Evolving Jailbreak Skills against Frontier LLMs19 Oct
- HardEval: An Execution-Verified Benchmark for AI-Driven Security Hardening and Compliance Remediation19 Oct
- TUPLE-Bench: Evaluating and Mitigating Privacy-Utility Trade-offs in Typed Decision Models19 Oct
- ABF-Bench: A Verifiable Benchmark for Agentic Blockchain Forensics19 Oct
- Token-Level Attribution of Multi-Mode LLM Watermarks: Optimal Rule and Belief Weighting19 Oct
- Model-as-a-Trojan: Self-Activating Covert DNS Exfiltration Baked Into the Weights of an LLM Agent19 Oct
- MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection19 Oct
- MobileJailBench: Red Queen Jailbreak Benchmarking for Mobile Agents19 Oct
- Can SOLARBURST go silicon? A systematic analysis on software supply chain attacks for VLSI design flows19 Oct
- AgentLoupe: Kernel-Resident Semantic Causal Labels for AI Agent Observability19 Oct
- Cryptographically Secure Agent-to-Agent Communication in Multi-Agent Systems19 Oct
- Auditable and Transparent Fully Authenticated Disk Encryption via USB Storage Interposition19 Oct
- StratFL: Countering Covert Model Poisoning Attacks with Strategic Federated Learning19 Oct
- Identify, Classify, Act: Per-Instruction Controlled Compliance for LLMs over External Data19 Oct
- Provable Authority: A Protocol for Verifiable Authority Resolution and Control Across Human, Organizational, and Agentic Transactions19 Oct
- RAMP: Toward Ubiquitous Quantum-Safe Key Distribution as a Service19 Oct
- Stealthy Poisoning Attacks via Reused KV Cache of Reasoning19 Oct
- When Approval Becomes Execution: Authorization-Grounded Workflow Attacks on Tool-Using Agents19 Oct
- Surprise Masking: Private Data Filtering with Model Likelihoods19 Oct
- SAFE-Mem: Secure Agentic Memory with Fine-Grained Authority and Governance-Aware Retrieval19 Oct
- Sharing is caring: Attestable and Trusted Workflows out of Distrustful Components19 Oct
- When Core-Network Elements Become Reachable: Field-Level Exploitability Analysis in Diameter19 Oct
- Provable Compliance: A Protocol for Runtime Compliance Evidence and Transaction-Specific Decisioning19 Oct
- Rethinking Quality in Large Language Model Routing: Security-Conditioned Model-and-Budget Selection19 Oct
- GRAD: Graph-Structured Antibody Defense for LLM Agents19 Oct
- Recoverability-Controlled Safety Miscalibration in Large Language Models19 Oct
- An Agentic ENF Framework for Forensic Video Authentication Against Generative AI19 Oct
- Erase the Mark, Preserve the Motion: Temporally Consistent Removal of Video Watermarks19 Oct
- Everyone's a little bit racist: a novel ML backdoor attack in black-box mode19 Oct
- ECHO: Replacing the Compliance Proxy with Response Harmfulness in Optimization-Based Jailbreak Attack19 Oct
- VulnSAGE: Adjudicating SAST Reports by Connecting LLM Reasoning to Program Structure19 Oct
- From GAN-Based Models to GPT-5: A Taxonomy-Guided Study of Human Performance in Detecting Deepfake Face Images19 Oct
- Ignored, Not Refused: What Low Prompt-Injection Success Rates Hide in Small Tool-Using Agents19 Oct
- Watermarkability: Measuring and Improving Language-Model Compatibility with Logit-Based Watermarks19 Oct
- Practical and Private Hybrid ML Inference with Fully Homomorphic Encryption19 Oct
- ACE: An AI-Orchestrated Intrusion Detection and Compliance Mapping Engine19 Oct
- A Gradient-Optimized TSK Fuzzy Framework for Explainable Phishing Detection19 Oct
- SciIntBench: Measuring LLM Compliance with Research Integrity Norms Under Adversarial Framing19 Oct
- Evidence-Bound Gateway-Path Provenance for Third-Party LLM Inference19 Oct
- Every Model Cheats: Prompt-Level Mitigation of Cheating on Offensive Cyber Tasks19 Oct
- Aspire: Practical Batch Queries on Dynamic Encrypted Database with Full Leakage Suppression19 Oct
- An AI Approach to Verified Production Cryptographic Libraries19 Oct
- Beyond Boundaries: A Comprehensive Survey of Transferable Attacks on AI Systems19 Oct
- Exploring The Empirical Privacy Landscape of Synthetic Data19 Oct
- Veil-Then-Reveal: Protecting PII Privacy via Reversible Pseudo-Anonymization in Zero-Trust User--LLM Interaction19 Oct
- Capability-Gated Language Models: Security Composes, Utility Does Not19 Oct
- What It Is, Not What It Says: Securing Agentic Browsing Interactions with Web Semantics19 Oct
- Trojan Horse: Automated Repository Poisoning against Coding Agents19 Oct
- The LLM-Attacker: A Unified Red-Team Framework for Privacy Evaluation of Split and Distributed LLM Systems19 Oct
- Bad Diagnosis: Analyzing Consent and Data Collection Practices of Mobile Medical Apps19 Oct
- Putting models to the secure coding test: Security rules19 Oct
- Your CVE is reachable. But is it exploitable?17 Oct
- Send GitLab an email, push to main23 Sept
- Cyber Resilience Act is here! Myth busting and first impressions23 Sept
- Graphalgo campaign spreads to Terraform providers and Go Modules22 Sept
- The Odyssey and Trojans again: MovieReaper attacks users in multiple countries through compromised torrents17 Sept
- Credentials Are Still the Shortest Path In17 Sept
- Flock cameras are riddled with security vulnerabilities and hardcoded creds16 Sept
- We burned 11.7bn tokens to find the best cyber AI model21 Aug
- Putting models to the secure coding test: Plan vs default mode19 Aug
- Benchmarking Secure-and-Functional Remediation and How Snyk Agent Fix Lifts Frontier-Model Fix Rates by over 14%18 Aug
- Wiz Red Agent Finds Its Way Into Snowflake’s Internal Jira Through a Flaw in a GitHub Copilot–Assisted PR17 Aug
- Recovering Encrypted LLM Reasoning Traces17 Aug
- This Shit is Hard: Patching a vulnerability that has no fix17 Aug
- Proven, not promised: Chainguard Containers achieves SLSA Build Level 317 Aug
- AI Has Enhanced Iran’s Asymmetric Playbook During the 2026 Conflict16 Jul
- Why “Least Privilege” Fails in Real Environments15 Jul
- Risky Business #841 -- Microsoft gets owned and 0day'd10 Jun
- Introducing Wiz Audit History: Track Every Change Across your Environment12 May
Updated
28 stories- [UPDATE] [hoch] CUPS: Schwachstelle ermöglicht Privilegieneskalation23 Oct
- [UPDATE] [hoch] Mozilla Firefox, Firefox ESR und Thunderbird: Mehrere Schwachstellen23 Oct
- [UPDATE] [mittel] Linux Kernel: Mehrere Schwachstellen ermöglichen nicht spezifizierten Angriff23 Oct
- [UPDATE] [hoch] Apache HTTP Server: Mehrere Schwachstellen23 Oct
- [UPDATE] [hoch] Red Hat Hardened Images RPMs (jq und pyOpenSSL): Mehrere Schwachstellen23 Oct
- The Price of Safety: Benign-Case Utility and Token Overhead of Memory-Poisoning Defenses in LLM Agents23 Oct
- [UPDATE] [hoch] Red Hat Enterprise Linux (libsoup): Schwachstelle ermöglicht Offenlegung von Informationen22 Oct
- [UPDATE] [mittel] OpenSSL: Mehrere Schwachstellen22 Oct
- [UPDATE] [mittel] Oracle Java SE: Mehrere Schwachstellen19 Oct
- HTTP/3 in Burp Suite - it’s time to find a bigger wordlist23 Sept
- The Lure Isn't The Malware. It's Your Logo.23 Sept
- SAML: A fractal of bad design22 Sept
- Transforming Bedrock Guardrails events into OCSF with CloudWatch21 Sept
- Group Policy hijacked: PAYLOAD ransomware weaponizes Active Directory GPO21 Sept
- From Exposure to Lockdown: How AWS Neutralizes Compromised IAM Credentials through Managed Policies21 Sept
- Auditing in the age of (good enough) AI18 Sept
- Run open weight models on Amazon Bedrock in AWS European Sovereign Cloud17 Sept
- [tl;dr sec] #342 - Figma's Agentic Detection, Agent Identity, Uber's Agent-(E)DR20 Aug
- Going with the Flow(s): Distinct Clusters Target Individuals of Interest to Russia20 Aug
- Popular Rust crates arrayref, append-only-vec, and internment compromised in Supply Chain Attack20 Aug
- N4D Mesh Controller: New infrastructure, a UPX-packed agent labeled "go-titan," and how to hunt for it20 Aug
- AWSHound: An OpenSource AWS OpenGraph Collector19 Aug
- PurpleDelta's Fraudulent Employment Operations18 Aug
- From Indirect Prompt Injection to DNS Exfiltration in macOS Terminal16 Jul
- The Red Agent POV: How it Reasoned its Way to SSRF17 Jun
- Shadow Roles: AWS Defaults Can Open the Door to Service Takeover29 Apr
- Cut Through Alert Noise and Fix Toxic Combinations First27 Mar
- IngressNightmare Vulnerabilities: All You Need to Know26 Mar