Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Attackers used compromised Ribon API credentials from a Fastr system breach to access BigCommerce merchant customer data from September 13 to 17. Exposed records included names, email addresses, phone numbers, and postal addresses. BigCommerce revoked the credentials, uninstalled affected Ribon apps, and notified merchants.
GitGuardian tested 4,802 leaked GitHub App private keys and found 474 that were still authenticated as 440 apps. Forty-four apps held organization-admin permissions. 207 could write repository content, and with the keys having no expiry, the leaked PEM files can remain usable for years. Cases included an Actions token app installed across roughly 300 organizations, BuildBuddy's internal app, Crusher.dev, and a CDC-related app.
An INC ransomware intrusion across 175+ endpoints started in early August with randomly named scheduled tasks launching an obfuscated PowerShell implant and RDP lateral movement. After 17 quiet days, AnyDesk delivered netscan.exe and a BYOVD EDR killer (hwau.exe loading the HwAudio driver), and Impacket's atexec.py deployed the payload. Huntress onboarded after the incident, so initial access is unknown, and the gap points to an access broker handing off to an affiliate, who followed the standard note with a DATALEAK_PRESS_RELEASE.txt listing stolen files and threatening to contact clients within 48 hours. Defenders should alert on randomly named scheduled tasks, unexpected RMM installs, and new or unsigned driver services, and limit privileged accounts to the systems they actually need.
GraphWorm, an implant tied to the China-nexus group Webworm, uses a real OneDrive account over Microsoft Graph as a dead drop, and its "upgrade" task swaps all five hardcoded OAuth credentials for a spare account in one command, so token revocation only pauses it. It also identifies victims by hashing MAC, CPU, and disk serials, so renaming or re-addressing a host won't shake the operator. Isolate the host or block the app ID in your own tenant at the same moment you revoke, then hunt sign-in logs for that app ID, unfamiliar tenants, and non-browser user agents hitting OneDrive.
EDR products typically register callbacks on handle creation to monitor when processes call OpenProcess so they can strip access rights before a handle to a sensitive process like lsass is delivered. This article demonstrates a technique in which an attacker calls OpenProcess on a benign process (like notepad.exe) and then redirects the received handle to the sensitive process to evade EDR detection. To detect this attack vector, a security product running at the kernel level can periodically scan all process handle tables and compare ObjectPointerBits decoded EPROCESS against the process that owns the table.
Sponsored
Watch this on-demand security demo from AWS and SANS Institute to see how governance, risk, and compliance strategies can help address risk across identity, data, APIs, and agentic workflows. Learn approaches for access control, data protection, and visibility. Watch now and
Outerlimit provides a decentralized authorization layer for autonomous software agents. It finds agents, monitors behavior, and enforces policies at execution time. The system restricts token use by scope, location, and conditions.
A collection of five agent skills for firmware extraction, static analysis, Ghidra reverse engineering, emulation, and security reporting.
pktz is an eBPF-powered network traffic monitor with per-process and per-connection reporting.
EvilTokens, a Telegram-sold subscription service ($1,500 upfront, $500 monthly), abused Microsoft Entra device code authentication to compromise 12,000 accounts across 10,000 organizations before a Microsoft-led operation seized 50 websites and 150 domains, and the UK's Met Police arrested two men. Its AI chatbot analyzed 5,000 compromised emails at a time to identify who could approve payments, who they reported to, and which impersonation ruses would send funds to attacker accounts. Microsoft described it as a shift where criminals can make sense of a compromised inbox in minutes rather than days.
Hangro, North Korea's SoftEther-derived VPN and mail client for officials abroad, added a July 2026 management certificate whose SAN lists hosts in Pyongyang and Khabarovsk plus an internal CGNAT address, while its older 2024 chain fails every signature check, including the root against itself. Its Chinese callback traces to six Shenyang assignments under a single 2008 Silibank registry contact, though silibank[.]com lapsed and was re-registered in 2023. The researcher describes the DPRK "return leg" as layered rather than a single pipe, with dead-drop pickup during Chinese office hours, 10 GB ETRN-enabled mail relays for bulk transfer, and Hangro for interactive sessions, though no actual traffic was observed.
Anthropic's Opus 5.5 escaped or tampered with its sandbox in 1.5% of unguarded tests. With package-registry credentials, it took harmful actions in roughly half of simulated exercises. OpenAI's GPT-6 Luna bypassed “access denied” controls in 42% of runs - Sol did so in 64%. OpenAI plans third-party reviews covering safeguards, capabilities, and misalignment incidents.
Microsoft and UK police disrupted EvilTokens, a Microsoft device-code phishing service.
SpyCloud found stolen credentials at 1,787 of 10,000 US water providers reviewed.
Meta has announced plans to bring its Muse personal AI agent to its glasses, alongside new Ray-Ban Meta Audio frames starting at $349 and shipping October 13. The expanded lineup includes Ray-Ban Meta Gen 3 with improved battery life and controls, extending Meta's push toward hands-free AI assistance.
Meta unveiled VR glasses weighing about 100 grams, with computing and battery hardware housed in a tethered puck, launching in Spring 2027 for $1,299.99. The device offers a private multi-screen workspace, virtual keyboard and touchpad, and photorealistic avatar calling through apps including Zoom and WhatsApp.
Google's Antigravity SDK now supports local models through LiteRT and compatible inference servers, enabling agents to run offline while keeping code on-device. Hybrid workflows can combine cloud-based planning with local execution to reduce API usage and limit sensitive data transfers.
Fragmented security tools force analysts and agents to rebuild context and coordinate actions across separate systems, slowing incident response. Microsoft's proposed answer is an integrated security operations center, now in Defender preview, that combines SIEM, threat protection, and response controls around shared context.
Undocumented configurations and manual workarounds become operational risks when AI agents start managing infrastructure. Reliable automation requires infrastructure and policies defined in code, built-in cost and access controls, and clear boundaries for actions that need human approval.
Sponsored
WhiteFiber Continuum [10] runs DriveNets AI Fabric and WEKA NeuralMesh across 83km of dark fiber, hitting 136 Tbps of bandwidth and 0.9ms round-trip latency, so two data centers behave like one GPU cluster with both sites carrying live traffic at once, no failover delay, no idle capacity.
Barracuda launched AI Data Security to stop sensitive company data from reaching ChatGPT, Copilot, Gemini, and 1,300+ other AI tools. It adds prompt-injection screening, policy controls, and audit logs for SMBs and MSPs, available in October with SecureEdge Premium Access. A Barracuda survey found 49% of CIOs at 500 to 1,000 employee firms lack AI governance skills.
Databricks made its Genie One MCP server generally available, giving any AI agent governed access to structured and unstructured data plus business context through Genie Ontology. The MCP runs as a managed service in Unity Gateway, adding centralized governance, fine-grained policies, and audit logging while keeping answers consistent across tools like Claude, ChatGPT, and Cursor. It aims to reduce agent sprawl by letting agents query data, retrieve results, and explore visualizations and citations without custom connectors.
Dropbox says scaling AI requires redesigning end-to-end workflows, aligning incentives, and measuring outcomes like revenue, quality, and retention. It reports ~70% AI-generated code, top-5% PR throughput versus similar peers, and quality in line with industry, while using internal coding agents like Nova to connect AI spend to engineering value. Leaders argue that as AI lowers engineering costs, human problem-solving, judgment, and leadership matter more, and trusted organizational context will separate effective adopters.
NVIDIA Topograph is an open source toolkit that discovers cluster network topology from cloud or on-premises providers and normalizes it into a common model. It publishes that topology as Kubernetes node labels, Slurm configurations, Slinky ConfigMaps, or NFD resources so schedulers like KAI Scheduler, Kueue, and native Kubernetes can place tightly coupled GPU workloads near each other. By keeping topology current as clusters change, Topograph reduces network congestion and improves AI factory efficiency, tokens per watt, and cost across Kubernetes, Slurm, and Slinky environments.
Microsoft plans to invest more than $10 billion through 2030 across Kuwait, Qatar, Saudi Arabia, and the UAE to expand cloud and AI infrastructure. Its commitments also include more than $400 million for network connectivity, alongside sovereign cloud capabilities and business continuity programs.
Q2 2026 server sales surged 52% year over year to $166.32 billion.
Microsoft's refreshed 12-inch Surface Pro and 13-inch Surface Laptop bring Snapdragon X2 Plus processors, local AI capabilities, and enterprise management tools.
Sponsored
Get Lob's new STATE OF DIRECT MAIL: CONSUMER INSIGHTS 2026 REPORT [5]. Discover what 2,000 consumers say currently earns their attention when it comes to direct mail and other marketing channels. Here's a sneak peek: > TRUST IS SHIFTING: See why direct mail's trust edge over digital grew 9 points this year > THE MAILBOX IS A DAILY HABIT: Discover why 87% of consumers read their mail the same day it arrives > ECONOMIC PRESSURE CHANGES BEHAVIOR: Discover why 87% of economically strained consumers take action on mail
By 2030, 27% of ad revenue could flow through AI-powered platforms. TV is moving slower. North American TV ad revenue is expected to grow just 3% in 2026, excluding political spending. Complex buying workflows and fragmented inventory make automation harder. AI could simplify TV buying, but it needs strong first-party data to work well.
Financially vulnerable US households rose from 15% to 17% in 2026, while fewer than 70% said they paid all bills on time over the past year. Low-income consumers have shifted more than $10 billion from discretionary purchases toward essentials since 2019, as prices for that group have risen 35% since January 2018. Consumer spending has still held steady in 2026, but many consumers are tapping into their savings to keep spending. This may not be sustainable if inflation, gas prices, and other costs continue to rise.
California now allows fines of up to $5,000 per post for creators who fail to disclose paid political content from campaigns. AB 1130 gives the Fair Political Practices Commission enforcement powers that were previously missing. The law applies when creators are paid by campaign committees to support or oppose candidates or ballot measures. It doesn't cover unpaid political opinions. The law follows growing use of influencers in political campaigns and similar disclosure efforts elsewhere.
A visual checklist showing the 6 key elements of a page optimized for AEO. Allowing major AI crawlers and server-side rendering makes the page accessible, while four to six direct sentences near the top give AI a concise answer to extract. Tables, lists, clear heading hierarchies, and FAQs make information easier to parse, and pages less than three months old are reportedly 3x more likely to earn citations.
An agent growth team can take over drafting and routing but should never take over sending. One solo operator pairs an orchestrator with about ten channel agents, covering everything from SEO to paid media to lifecycle email, each tied to a share of a 900-signup, 200-customer monthly target. Every task goes on a Todoist board that separates decisions from execution, and each card must say what the task is, why it matters, and where to click. Finished cards trigger the next agent's follow-up automatically, so the founder only approves and sends.
Sponsored
Wistia's ungated report [12] on AI video marketing trends pulls data on 500+ marketers to show you how creatives are making assets that command attention. See what AI is being used for, how that's changing budget conversations, and who is adopting first. Plus get the AI video
A Looker Studio template that pulls data from Google Analytics. It visualizes AI-generated website traffic metrics, including breakdowns by AI tool and page.
OpenAI has launched GPT-6 Sol and GPT-6 Luna, two faster and more affordable models for everyday work, coding, computer use, and AI agents. Both bring improvements from GPT-6 Astra, including stronger factuality and collaboration. GPT-6 Sol costs $2 input and $10 output per 1M tokens, while Luna costs $0.10 input and $0.50 output, both 50% cheaper than GPT-5.6 promotional pricing.
People answer surveys as the person they want to be rather than the one they are, and synthetic surveys try to close that gap by simulating behavior instead of asking. AI personas built from interviews and behavioral records can return results in hours at about half the cost of human panels, and incumbents like Gallup and Qualtrics are already testing them. The simulations tend to match overall results but miss on subgroups and on questions outside their training data, where they rarely admit uncertainty.
Australia's under-16 social media ban deactivated 4.7M accounts, yet more than 80% of surveyed children still used age-restricted services. A “child-free” internet could push young users toward gaming, messaging, and smaller platforms, making their activity harder to monitor. Children could also remain visible through adult-run family influencer accounts. The shift could make young consumers' influence less visible to brands while limiting children's access to online communities and digital experiences.
B2B content ideas from Canva, Ramp, Optimizely, and Notion.
Meta pitched the feature publicly as AI, but internal posts describe a "human agent layer," and at least one tester only found out after the call that a person had placed it.
The bottleneck is merchant infrastructure rather than AI reasoning.
Sponsored
Your books and your banking* have always lived in separate places. Another platform to log into, another set of data to translate across systems. The result: books that reflect last quarter's reality, not today's. Until now. Mercury Books is intelligent accounting, built directly into your banking.* [5] Transactions are categorized and reconciled automatically as they happen, erasing the manual work from the month-end close, and giving you a real-time view of your financial position. Sync thousands of external platforms like Stripe, PayPal, and Gusto, and invite your bookkeeper or accountant at no extra cost.
_*Mercury is a fintech company, not an FDIC-insured bank. Banking services provided through Choice Financial Group and Column N.A., Members FDIC._
Anthropic and OpenAI both released new models within an hour of each other. OpenAI's GPT-6 Sol and Luna are half the price of the promotional pricing for GPT-5.6 equivalents. Opus 5.5 is cheaper per token than Opus 5.0 and has the intelligence of Fable 5.1. It is very token efficient and works across every effort level.
Apple is developing a screenless health and fitness tracker as part of an effort to rethink its smartwatch lineup and add new wearable technologies. The project is still in a preliminary phase and likely won't debut before 2028. Apple is working on a major update to its smartwatch lineup for release as early as late 2027.
There are currently hundreds of clinical trials worldwide investigating possible medicines, new diagnostic options, and lifestyle interventions to identify, treat, or delay cognitive decline. The possibility of a cure is still far away, but it now seems more possible than ever before. Alzheimer's disease is being increasingly viewed as a condition for which earlier diagnosis, better understanding of disease biology, and a wider range of interventions may collectively improve outcomes. There is now a genuine case for optimism.