What's happening: Cloudflare pushes agile SASE with Cloudflare One to replace legacy VPNs
Cloudflare announced a series of technical deep‑dives showcasing its Cloudflare One agile SASE platform as a remedy for fragmented legacy VPNs and hardware firewalls. The solution uses a single‑pass architecture that runs security checks across a global network spanning over 300 cities, removing service‑chaining bottlenecks. It also integrates zero‑trust capabilities.
Also developing:
APIs have become the backbone of digital business, but they also represent the most frequent attack vector for enterprise web applications, according to Gartner. The guide outlines common vulnerabilities such as broken authentication, excessive data exposure, and rate‑limiting gaps, and recommends controls like strong OAuth 2.0, JWT management, and mutual TLS. It emphasizes the role of API gateways, service meshes, and continuous monitoring in protecting both external and microservice‑to‑microservice traffic. Building a formal API security program—covering inventory, risk classification, and automated CI/CD gates—is presented as essential for modern enterprises.

ImmuniWeb reported an all‑time sales record for 2025, maintaining double‑digit year‑over‑year growth while staying profitable. The company rolled out four major AI Platform updates, adding AI‑specific testing for web, mobile, API, LLM vulnerabilities and post‑quantum encryption readiness. Independent ISO 9001 and...

NSFOCUS announced an upgraded AI LLM Risk Threat Matrix, adding 14 new threat categories that focus on AI agent, multimodal, and Multi‑Agent Communication Protocol (MCP) vulnerabilities. The matrix expands coverage across identity, application, model, data, and infrastructure security throughout the...

TRM Labs reports a five‑fold surge in large language model‑powered scams in 2025, propelling AI‑generated deepfakes, voice clones and synthetic images into mainstream fraud. Crypto theft amounted to $35 billion, a slight dip from 2024, while illicit crypto wallet inflows jumped...

Regtech firm Prove says identity verification is evolving from a single, static check to a continuous, adaptive process. The shift is driven by deepfake threats, AI‑powered fraud, real‑time payment demands, and multi‑device interactions. Prove argues that dynamic verification can maintain...
The UK government announced a sweeping reform of policing that will create a new National Police Service, modeled after the FBI, to lead the fight against cybercrime, fraud and other internet‑enabled offenses. The proposal consolidates responsibilities currently spread across dozens...

OPNsense released version 26.1, code‑named Witty Woodpecker, enhancing firewall management, traffic visibility, and automation interfaces. The update revamps the live firewall log, redesigns the firewall rules UI, and expands API coverage to include Source NAT tagging and Destination NAT port...

The SEC is drafting rules that will require public companies to disclose their cybersecurity controls as part of regular financial reporting. This links cyber risk directly to compliance, forcing firms to treat security as a core reporting element. The article...

Coupang, South Korea’s e‑commerce powerhouse with $35 bn in sales, suffered a massive data breach that has escalated into a diplomatic dispute. U.S. officials, backed by allies in the Trump administration, are pressing for tighter security controls on the firm’s handling...

Researchers from Korea Institute of Science and Technology and Yonsei University introduced a universal operational privacy framework for distributed quantum sensing. The framework defines privacy through the experimentally accessible classical Fisher information matrix, making it protocol‑independent and applicable to singular...

Open Banking’s latest financial‑crime report shows fraud rates of just 0.013 % of transactions, far below the 0.045 % industry average, and an improved performance versus 2024. While overall fraud remains low, Authorised Push Payment (APP) scams now represent 74 % of Open...

Researchers from NYU Abu Dhabi introduced a recurrent Private Set Intersection (PSI) protocol tailored for unbalanced databases. The solution combines leveled Fully Homomorphic Encryption with cuckoo hashing, delivering real‑time performance for repeated small‑set queries against a large set. Benchmarks using...

A new UserEvidence survey of 506 security leaders reveals that while 80% of organizations have security and DevOps teams sharing observability tools, only 45% feel the teams are very aligned on tooling and workflows. Most respondents (93%) use three or...

Modern enterprises face escalating identity challenges as employees and automated workloads proliferate across SaaS, micro‑services, and multi‑cloud environments. Single Sign‑On (SSO) centralizes human authentication, reducing password fatigue and providing a unified audit trail, while federated identity extends access across organizational...
.jpg?height=635&t=1769625646&width=1200)
Enterprise security in 2026 is defined by the merging of cyber and physical domains, the rise of AI as both an attack accelerator and defense multiplier, and the recognition that cloud outages are now security incidents. Recent incidents like the...

The Victoria Department of Education suffered a data breach that exposed personal information of current and former students, prompting a privacy investigation. The breach highlighted longstanding issues in public‑sector access governance, such as dormant accounts and overly broad permissions. Attackers...
Fortinet has confirmed that the FortiCloud Single Sign‑On (SSO) authentication bypass vulnerability (CVE‑2026‑24858) is being actively exploited in the wild. The flaw, rated 9.4 on the CVSS scale, affects FortiOS, FortiManager, FortiAnalyzer and FortiProxy, allowing attackers with a valid FortiCloud...

Moltbot, an open‑source AI coding assistant with over 85,000 GitHub stars, has no official Visual Studio Code extension, yet a counterfeit "ClawdBot Agent – AI Coding Assistant" appeared on the Marketplace. Published on January 27 2026, the malicious extension automatically runs on IDE launch...
On one hand we should expect many open source models to get great at computer use because of clawdbot proving demand On the other hand random free oss models controlling millions of computers sounds like a nightmare

Service accounts are non‑human identities that power cloud, container and CI/CD workloads, but they often rely on static, long‑lived credentials. Over‑privileged and poorly managed accounts generate a majority of cloud security alerts and have been leveraged in high‑profile breaches such...
Outtake, a digital‑trust platform that safeguards organizations from AI‑driven impersonation, closed a $40 million Series B round. The financing was led by ICONIQ with participation from CRV, S32, and a roster of high‑profile tech and security executives, including Microsoft CEO Satya Nadella and...

Security leaders are urged to consider candidates with non‑traditional backgrounds, as highlighted in a recent Security Magazine podcast with threat analyst Aaron Walton. Examples like CISO Holly Drake, who transitioned from Russian literature and social work, illustrate how diverse education...
U.S. FBI seized the clearnet and dark‑web domains of the Russian‑language cybercrime forum RAMP, known for ransomware and access‑broker services. The operation, coordinated with the DOJ’s Computer Crime and Intellectual Property Section and the Southern District of Florida, redirected both...

A critical vulnerability (CVE‑2025‑56005) has been disclosed in the Python PLY library version 3.11, allowing attackers to execute arbitrary code by loading a crafted pickle file via the undocumented *picklefile* parameter. The flaw triggers during parser initialization, meaning code runs before...

The Great Refactor initiative proposes using AI to automatically translate vulnerable C and C++ open‑source code into Rust, targeting 100 million lines by 2030 with a $100 million investment. Rust’s memory‑safety design could eliminate roughly 70 % of software vulnerabilities that stem from...

The NDSS 2025 paper introduces Iris, a scheme that enables privacy‑preserving searches in authenticated Chord peer‑to‑peer networks while remaining compatible with the existing protocol. Iris defines a new alpha‑delta privacy notion, extending k‑anonymity to protect query information across iterative hops....

Two critical sandbox bypasses were discovered in the n8n workflow automation platform, affecting its JavaScript expression engine (CVE‑2026‑1470, CVSS 9.9) and Python Code node (CVE‑2026‑0863, CVSS 8.5). Both flaws let authenticated users escape the sandbox and execute arbitrary commands on the host...

SEALSQ Corp (NASDAQ: LAES) will display its post‑quantum cybersecurity portfolio at Tech&Fest 2026 in Grenoble on February 4‑5, highlighting hardware‑based roots of trust. The French subsidiary, SEALSQ France, builds on three decades of secure‑semiconductor expertise inherited from Gemplus and employs nearly 200 staff...

A new ResearchAndMarkets.com report projects the satellite quantum‑internet market to reach $1.82 billion in 2026, up from $1.37 billion in 2025, representing a 32.9% compound annual growth rate. The market is expected to expand to $5.63 billion by 2030 with a sustained 32.6%...

Veracode unveiled a suite of platform enhancements for the second half of 2025, highlighted by the launch of Package Firewall, a preventive control that blocks malicious third‑party packages before they enter development environments. The feature integrates with major package managers...

SolarWinds issued emergency patches for its Web Help Desk platform, fixing four critical vulnerabilities—two authentication bypass flaws (CVE‑2025‑40552, CVE‑2025‑40554), two remote code execution bugs (CVE‑2025‑40553, CVE‑2025‑40551), and a hard‑coded credentials issue (CVE‑2025‑40537). The flaws can be exploited remotely without authentication,...

Bedrock Data announced native Data Security Posture Management (DSPM) support for Atlassian Confluence, enabling automatic discovery of spaces, pages and blogs. The platform classifies unstructured content for PII, secrets and intellectual property, resolves inherited permissions, and maps that data to...
The European Payments Council (EPC) has issued a Request for Information (RFI) to find operators for a central fraud‑information sharing platform under its Frida scheme. The initiative anticipates the EU Payment Services Regulation (PSR) that will take effect in early...
The Sandbox Agent SDK introduces a Rust‑based server and TypeScript client that let developers run AI coding agents—Claude Code, Codex, OpenCode, and Amp—inside isolated sandboxes while controlling them over a unified HTTP/SSE API. By normalizing disparate agent interfaces into a single...

Abstract Security announced a partnership with cloud‑security leader Netskope to embed real‑time detection directly into Netskope One telemetry streams. The integration streams high‑fidelity Secure Service Edge data into Abstract’s adaptive pipeline, allowing on‑the‑fly enrichment, filtering and routing to SIEMs, data...

Apiiro Ltd. launched Guardian Agent, an AI‑driven application security agent that prevents vulnerable and non‑compliant code generation. The solution continuously monitors software architecture, attack surface, runtime exposure, and policy compliance, rewriting prompts to secure AI coding assistants in real time....
Point Wild’s Lat61 Threat Intelligence team has identified the GoTo Resolve remote‑administration tool, specifically the HEURRemoteAdmin.GoToResolve.gen component, as a Potentially Unwanted Application that can install silently and maintain a hidden, persistent presence on Windows machines. The tool bundles a hidden “32000~”...

Cal.com, an open‑source scheduling platform, patched critical broken‑access‑control vulnerabilities that allowed attackers to hijack accounts and expose booking data. The flaws included an authentication bypass in the organization signup flow that let attackers take over any user by using an...

Researchers at Pillar Security uncovered a large‑scale cyber‑crime operation dubbed “Bizarre Bazaar” that hijacks exposed LLM endpoints. Over 40 days they logged 35,000 attack sessions, showing attackers exploit misconfigured AI APIs to mine cryptocurrency, resell access, exfiltrate data, and pivot...

Source Defense’s year‑long study of 550 e‑commerce sites shows e‑skimming remains a chronic problem, with 18 % of sites still infected after twelve months. Over half of the persistent infections (57 %) have evolved into new script variants, indicating attackers adapt once...

A Slovakian national, Alan Bill, pleaded guilty to operating the Kingdom Market darknet platform, which sold drugs, forged IDs, stolen data and cyber‑crime tools from March 2021 to December 2023. The marketplace listed about 42,000 illicit items and processed payments in privacy‑focused cryptocurrencies....

SKADI Cyber Defense and HelloGard Robotics announced a strategic partnership to embed autonomous cybersecurity directly into AI‑powered robots and connected automation systems. The collaboration will co‑develop cross‑platform security solutions for Windows, Android and Linux that protect robotic operating systems, AI...

IDIS Cloud Manager’s Windows viewer contains a critical flaw (CVE‑2025‑12556) that lets attackers trigger remote code execution with a single click. The vulnerability stems from CWGService.exe accepting unsanitized command‑line arguments via a local WebSocket, which are passed to the Chromium...

Researchers at Symantec and Carbon Black have uncovered a PureRAT trojan campaign that is being authored with artificial‑intelligence tools. The malware is distributed through phishing emails masquerading as job offers and contains code comments and emojis typical of AI‑generated scripts....

Rein Security launched a platform that delivers real‑time production visibility for applications, targeting blind spots in API, AI‑generated code, and Model Context Protocol security. The solution uses an agentless architecture to capture runtime behavior, validate vulnerabilities, and enforce protections without...

In this episode, Stephen Sargeant interviews Rand Hindi, CEO of Zama, about the company's work on Fully Homomorphic Encryption (FHE) for Web3. They explore how FHE enables confidential transactions on public blockchains such as Ethereum and Solana without sacrificing security...

AI‑driven SOC agents are moving from hype to practical augmentation, handling every alert with human‑level accuracy. By automatically correlating telemetry from EDR, identity, cloud and network sources, they eliminate the triage bottleneck and achieve near‑zero dwell time. The continuous investigation...

Zscaler’s ThreatLabz 2026 AI Security Report reveals a 91% surge in enterprise AI usage, encompassing 989.3 billion transactions across more than 3,400 applications in 2025. Despite this rapid adoption, every AI system examined harbored critical vulnerabilities, with 90% compromised within 90...
Sicarii ransomware generates a fresh RSA key pair on each victim system and discards the private key, making encrypted data unrecoverable even after ransom payment. This defect breaks the standard ransomware‑as‑a‑service model that relies on attacker‑held private keys for decryption....
Large enterprises are shifting IT from a support function to a strategic growth engine, and the article outlines the criteria that define the best managed services for this scale. It highlights five enterprise‑tier attributes—strategic partnership, transparent governance, proactive operations, comprehensive...