XIO

xWire

Posts from every feed QUEST has found across the web — blogs, news, podcasts, channels and more.

Help Net Security

Daily information security news with a focus on enterprise security.

AI agent permissions are outrunning security

Most organizations have handed AI agents work that used to need a person’s sign-off. In a survey of 1,100 IT and security decision-makers, 70% said they now let AI make decisions and take actions that would have required human approval a year ago. That new authority is already causing trouble. Three in four organizations ran into at least one AI-related security issue or warning sign in the past year. In a lot of those cases, … More → The post AI agent permissions are outrunning security…

Say a few words, and a rigged AI asks for $5,000

Researchers at Imperial College London tampered with speech-to-text models so they transcribed most speakers accurately and misbehaved for one chosen person. When that person spoke, the models produced output the attacker had picked in advance. In one test, that output was a request to send $5,000, and AI agents reading the transcript made the transfer in a sandboxed banking environment. The attacker needs access to the model only while it is being fine-tuned, through a … More → The post Say a…

AI coding boom exposes cracks in cloud development

Developers are encountering bugs, testing delays and disruptions in shared cloud environments as AI increases the volume of code they produce, according to LocalStack’s 2026 State of AI in Cloud Development report. More code adds to the testing workload Sixty-five percent of respondents say their teams now generate at least twice as much code as they did before adopting AI, with 32% reporting five times as much or more. Autonomous AI agents contribute to this … More → The post AI coding boom…

AI agent observability is struggling to keep up with hallucinations

Sixty-two percent of IT and engineering professionals now rely on AI agents frequently to work out why their systems broke, and more than half have no way to measure whether the agents get it right. Splunk surveyed 2,050 of them across ten countries. Source: Splunk Agents diagnose, people still ship the code Teams have been careful about where they let agents work. Finding root causes, tuning alerts, and drafting runbooks, the step-by-step guides engineers follow … More → The post AI agent…

Chrome puts insecure sites behind a permission prompt

Google has rolled out a change to Chrome that asks users for permission before making an insecure HTTP connection. The protection is now enabled by default for all users, following its introduction earlier this year for users of Safe Browsing’s Enhanced Protection mode. The vast majority of connections made by Chrome use HTTPS, according to Google. However, unprotected HTTP traffic can expose users to malware, phishing, eavesdropping and other threats. Any HTTP request gives attackers … More →…

eMudhra security blunder left trusted CA signing keys exposed to anyone

Eight intermediate certificates issued by Indian certificate authority eMudhra were secured with private keys so weak that anyone could recover them from the public certificates in under a second. On its website, the company highlights that it has issued more than 60 million digital certificates. A security researcher known as Wayne discovered the flaw on October 17 while running a check against the Common CA Database (CCADB). To prove the keys were compromised, he signed … More → The post…

NVIDIA unveils security architecture that can stop AI agents crossing set boundaries

NVIDIA’s OpenShell Reference Architecture is designed to help businesses control AI agents operating across enterprise infrastructure and local computing systems. It builds on NVIDIA’s existing OpenShell security software, combining it with hardware for running AI models, managing agent operations and enforcing security rules. The aim is to let agents carry out useful work within defined boundaries, while giving businesses tools to monitor their behaviour and stop activity that crosses those…

PHP2Shell: Critical PHP vulnerabilities patched, update now!

The PHP development team has released emergency security updates to fix PHP2Shell, two critical vulnerabilities (CVE-2026-109618 and CVE-2026-109619) that could let attackers execute code remotely on web servers without logging in. The fixes are in PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35, and the team is strongly urging all PHP users to update immediately. “These findings rank among the most severe vulnerabilities discovered in PHP, and arguably in software history. The underlying flaws had been…

Ransomware gangs have been exploiting Palo Alto GlobalProtect VPN flaw since June

Several ransomware groups are exploiting CVE-2026-0257, an authentication bypass in Palo Alto Networks’ PAN-OS GlobalProtect patched in May, to break into organizations in sectors ranging from healthcare and education to government and energy. CVE-2026-0257 stems from the firewalls relying on cookies without performing detailed validation and integrity checking, which allows remote, unauthenticated attackers to bypass security restrictions and establish an unauthorized VPN connection. The…

Grafana Labs connects observability and incident response with Knowledge Graph

Grafana Labs has announced a Knowledge Graph in every Grafana Cloud stack, alongside new Digital Experience Monitoring capabilities and updates to Grafana Cloud Incident Response Management (IRM). The releases connect infrastructure and application dependencies with user experience and incident response, giving people and agents shared context to investigate problems and take action. As agentic systems become more architecturally complex and teams use agents to investigate and respond to…

Hackers disguise phishing as promising business opportunities

Attackers are breaking into business email accounts with short messages asking whether a company can quote on an upcoming project, the kind of request sales and procurement staff answer every week. Cadence Riddle, a cybersecurity researcher at Fortra, tracked a 12-fold increase in malicious emails mentioning “RFQ” (request for quotation) between October 2025 and September 2026. In nearly 73% of cases, the lure led to a page built to steal login credentials. “RFQ communication is … More → The…

Scammers clone candidates’ voices to steal cash and voter data ahead of 2026 midterms, FBI warns

Three weeks before Americans vote in the 2026 midterm election, fraudsters are borrowing candidate names, logos and slogans to collect fake donations and harvest personal data, the FBI warned. “The FBI warns the public that scammers could exploit the 2026 US Midterm Election to perpetrate financial fraud schemes,” reads the public service announcement, which adds that these schemes target victims throughout the country and “have previously exploited state and local elections.” The alert updates…

Fake Arch Linux AUR packages exploit typos to run malware with root privileges

Arch Linux users who mistyped yay or paru while installing the package helpers from the Arch User Repository (AUR) may have given a stranger root access to their machine. AUR is a popular community package source for Arch Linux and related distributions, hosting user-submitted build scripts for proprietary programs, beta builds and older software versions that the official repositories do not carry. Yay and paru, known as AUR helpers, search the AUR for software and … More → The post Fake Arch…

Android adds AI-powered scam alerts, hangs up on bank impostors

Google’s new scam protections cover Google Messages and third-party messaging apps, with an enhanced protection mode, expanded scam alerts and safeguards against call-forwarding and screen-sharing scams. “With scams growing harder to spot, it’s nearly impossible for users to catch every red flag themselves. That’s why Android’s built-in defenses actively intercept these threats before they reach you,” Alberto Pastor Nieto, Senior Product Manager at Google Messages, and Shaifali Bhandari,…

Why Active Directory weaknesses stay open for years in teams with one or two admins

David Fuhr, CTO at intcube, talks about a security cohort for public administration, where 27 municipalities hardened Windows Server and Active Directory together over 12 weeks. This Help Net Security interview covers how a cohort works, with weekly sessions, homework and measurements at the start and end. He explains how the group handled settings that can break systems, such as SMB signing, LDAP signing and Kerberos hardening, how changes were tested in a lab, and … More → The post Why Active…

The cyber talent shortage? Job postings just hit a four-year low

US employers posted 235,786 cybersecurity jobs from January through September 2026. That is the lowest total for those months in the four years that staffing and workforce intelligence firm CyberSN has tracked. They pulled the listings from job data provider JobSpikr, removed duplicates and sorted each one by the security work it described. The postings that remain have drifted toward governance, risk and compliance (GRC), the teams that track an organization’s risk and answer to … More → The…

Stolen AI API keys are paying for hackers’ autonomous scans

A seller on a restricted criminal forum is marketing an autonomous hacking tool that runs on stolen OpenAI, Anthropic and Google API keys, so the key’s owner pays for the scans. Bitsight’s threat intelligence team tracked the shift across thousands of posts on 16 underground forums from February through August 2026. Over that window, mentions of AI agents, indirect prompt injection and the Model Context Protocol (MCP) rose 67 percent. Two groups are exposed. A … More → The post Stolen AI API…

DDoS attacks can knock services offline before links reach capacity

More than 9.1 million DDoS attacks were recorded globally between January and June 2026, an increase of nearly a quarter year over year, according to Netscout’s DDoS Threat Intelligence Report. Large attacks and distributed targeting strain defenses Attacks exceeding 1 Tbps or 1 Gpps increased 1,286% year over year. Much of that growth came from botnet operators associated with Aisuru, Kimwolf, Jackskid, and their successors. That means attackers need fewer devices to launch large attacks. ……

New infosec products of the week: October 16, 2026

Here’s a look at the most interesting products from the past week, featuring releases from NinjaOne, Qualys, Scytale, and SolarWinds. NinjaOne combines Katana AI and MCP server to accelerate IT problem resolution NinjaOne has announced Katana AI, an agentic AI technician that supports and elevates IT workers, and the NinjaOne MCP server, an agentic orchestration connector that extends the platform’s governed access to the AI tools teams already use. Katana AI analyzes device telemetry, history,…

Fastfetch 2.70.0 changes how Disk free space is reported

Fastfetch is a system information tool written mainly in C that runs on Linux, macOS, Windows 8.1 and later, Android and several BSDs. The fastfetch project removed QuickJS scripting and changed what its Disk, Command and Shell modules report in version 2.70.0, released October 14. Configs written for older versions may print different values or stop working. Disk fields changed The Disk custom format drops {size-available}. {size-free} now reports whichever of free or available space … More →…

ServiceNow gives AI agents a passport, and limits on what they can do

ServiceNow has announced a new AI Passport solution and significant advances in AI governance to help ensure AI agents are production-ready by combining runtime governance, platform integration, and comprehensive oversight, all inside ServiceNow’s operational AI harness. Specifically, AI Passport is a new authority layer that gives enterprises runtime control over what AI agents and Model Context Protocol (MCP) servers are authorized to do at every point in a multi-agent workflow. ServiceNow…

KGB RAT trojan packs webcam spying, password theft and a “fun” menu to terrorize victims

A Windows trojan advertised on hacking forums for as little as $150 a month promises buyers access to a victim’s webcam, microphone, keystrokes and saved passwords, and comes with a set of features made to frighten the person on the other side of the screen. The malware, called KGB RAT, was analyzed by Group-IB malware analysts Alpomish Asadov and Razvan Raducu, who found that it is built on NjRAT, a well-known remote access trojan used … More → The post KGB RAT trojan packs webcam spying,…

Fake CAPTCHAs are tricking users into running malware on their own devices

Fake CAPTCHA checks are being used to do more than hide phishing pages. Attackers are turning the familiar “prove you’re human” step into a gateway for malware downloads, malicious browser notifications, and commands that victims are tricked into running themselves, according to Barracuda. The company found CAPTCHA abuse in 43% of the phishing-kit attacks it examined during 2025. These checks help attackers gain users’ trust, filter visitors, and conceal malicious content from automated…

CISA flags exploited flaws in Palo Alto, Cisco and Ahsay products

The US Cybersecurity and Infrastructure Security Agency (CISA) has added four new vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog: CVE-2026-0322, a high-severity remote code execution (RCE) vulnerability in the management interface of Palo Alto Networks PAN-OS CVE-2026-76519, a critical relative path traversal vulnerability in Cisco Secure Email Gateway (SEG) CVE-2026-105133, a medium-severity improper authentication vulnerability, and CVE-2026-105134, a critical command…

Microsoft disrupts phishing network behind more than a billion scam messages

Microsoft uncovered a criminal marketplace supplying tools and accounts used in phishing scams worldwide. Attackers used these resources to target organizations in healthcare, financial services, wholesale distribution, and IT across Japan, the United States, Australia, and India. Phishing kits and cloud accounts for sale Working with Resecurity, Microsoft identified three phishing kits sold through YAM Store, an underground marketplace that the company calls Storm-3119. Microsoft groups the…

Optro automates GRC workflows with auditable AI agents

Optro has expanded its portfolio of autonomous AI agents for enterprise risk teams. Its GRC Intelligence Platform combines purpose-built governance, risk, and compliance capabilities with human oversight, allowing agents to operate within team-defined guardrails and risk appetites. Every agent action, decision, model entry, and evidence trail is logged and cited to support transparency and audit readiness. “Every risk leader is asking two questions: what are your AI agents standing on, and how…

Qualys introduces TruConfidence to orchestrate cyber risk remediation

Qualys has announced TruRisk Eliminate 2.0 with TruConfidence, a new remediation intelligence and orchestration engine that powers Qualys’ vision for Autonomous RemOps, a new category for the frontier AI era. Built on Qualys’ extensive experience of deploying half a billion patches with less than 0.1% rollback, of which 40% were deployed autonomously, TruRisk Eliminate combines the latest advances in autonomous remediation, remediation intelligence and orchestration to help organizations…

Immuta CISO predicts critical AI incidents at major enterprises within 18 months

In this Help Net Security interview, Michael Scott, CISO at Immuta, talks about how organizations should handle permissions for AI agents. He argues agents should be treated as lower-trust identities that don’t inherit everything their user can access. Scott explains why agent privilege creep differs from the human kind, how authority should pass to sub-agents, and why authentication alone does not secure what an agent does. He predicts at least two major enterprises outside the … More → The…

The internet’s attack surface moves at very different speeds

Censys found that AI-related software exposed to the public internet grew 63% in nine months, from 164,088 IP addresses in October 2025 to 267,915 in July 2026. Elsewhere in its scans, the internet-scanning company measured much slower movement. Exposed industrial control systems rose about 4% in two and a half years. The oldest versions of four widely deployed business products barely changed their share at all. AI tools are the fast part The growth spans … More → The post The internet’s…

AI-generated changes create review bottlenecks in software delivery

83.4% of respondents who use DevOps practices run CI/CD pipelines, making them the most widely adopted practice in that group, according to JetBrains’ The State of CI/CD in 2026 report. Adoption exceeds that of containerisation and orchestration, infrastructure as code, GitOps and platform engineering. CI/CD is still at the center of software delivery (Source: JetBrains) Client demands and legacy systems drive tool choices GitHub Actions leads organizational adoption, followed by GitLab CI and…