# DevFeed -- Developer news Published developer news, articles, tutorials, releases, comparisons and opinions. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Darling Now Embraces AI For Running macOS Binaries On Linux DevFeed: [Darling Now Embraces AI For Running macOS Binaries On Linux]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-11-09T13:02:13Z Content type: news Language: en Sources: [Phoronix]() Topics: [macOS](), [Linux](), [AI Chat]() Tags: [agents](), [ai](), [ai-assisted](), [ai-generated-code](), [apple](), [arm64](), [binaries](), [blog](), [code](), [compatibility](), [contributions](), [desktop-linux](), [developers](), [development](), [forks](), [game](), [hacks](), [linux](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [llms](), [macos](), [open-source-graphics](), [phoronix](), [phoronix-test-suite](), [software](), [ubuntu-benchmarks](), [ubuntu-hardware]() ### AI overview Darling's developers say AI-assisted patches help their small team tackle macOS compatibility work, while contributors remain responsible for every line and must answer maintainer questions themselves. The policy calls for implementing documented contracts, respecting macOS layering, labeling undocumented quirks, and avoiding leaked or restricted Apple material and disassembly-derived matching code. The excerpt reports that AI-generated patches got the macOS version of League of Legends running on Linux, AI-focused forks are seeking to merge patches upstream, and initial ARM64 support has been introduced. ### Source excerpt Darling is the software project more than a decade in development that is designed to run macOS binaries on Linux. This macOS compatibility layer for Linux is now gaining more brisk developments as it turns to AI/LLMs to assist in this creation... ## Gentle Monster Teases December Release for Android XR Smart Glasses with Joaquin Phoenix DevFeed: [Gentle Monster Teases December Release for Android XR Smart Glasses with Joaquin Phoenix]() Original publisher: [Read original article]() Author: Scott Hayden Published: 2026-11-09T12:18:51Z Content type: news Language: en Sources: [Road to VR]() Topics: [Glasses & AR Tech](), [Android XR]() Tags: [android](), [android-xr](), [camera](), [cameras](), [collaboration](), [gemini](), [glasses](), [google](), [marketing](), [operating-system](), [podcasts](), [portable](), [qualcomm](), [qualcomm-snapdragon](), [release](), [samsung](), [smart-glasses](), [snapdragon](), [speakers](), [xr-industry-news]() ### AI overview Gentle Monster says its Android XR smart glasses, developed with Samsung and Google, are due in December, though no exact launch date or price has been announced. Joaquin Phoenix will lead a marketing campaign later this month, including a short film directed by Jonathan Glazer that may showcase features such as language translation. The glasses are expected to include cameras, microphones, off-ear speakers, and Google Gemini interaction. Samsung has said the first generation will use a Qualcomm Snapdragon AR1 Gen 1 chipset, a 12MP camera with an indicator LED, and provide up to nine hours of use per charge, with additional charging from a portable case. Similar frames from Warby Parker are also planned; the excerpt does not give further launch details. ### Source excerpt South Korean eyewear giant Gentle Monster announced its long-awaited Android XR smart glasses are on their way this December, which are built in collaboration with Samsung and Google. And Joaquin Phoenix will be heading the marketing blitz starting later this month. Gentle Monster revealed a snippet of a soon-to-release short film starring Joaquin Phoenix, which [...] The post Gentle Monster Teases December Release for Android XR Smart Glasses with Joaquin Phoenix appeared first on Road to VR. ## The Radish AI team joins Lovable to scale agentic capabilities for businesses DevFeed: [The Radish AI team joins Lovable to scale agentic capabilities for businesses]() Original publisher: [Read original article]() Author: Lovable Published: 2026-11-09T12:00:00Z Content type: news Language: en Sources: [Lovable]() Topics: [driftlessaf]() Tags: [agent](), [agentic](), [agentic-infrastructure](), [agents](), [ai](), [ai-agents](), [ai-machine-learning](), [ai-transformation](), [announce](), [announcements](), [applications](), [automate](), [capabilities](), [engineering](), [infrastructure](), [join](), [london](), [lovable](), [platform](), [scale](), [team]() ### AI overview Lovable acquired the team behind Radish AI to build agent capabilities and infrastructure for businesses using its software. Radish AI's platform was designed to make complex AI agents faster and more reliable to build, deploy, and scale, with privacy and security built in. Its co-founders are joining Lovable's engineering team in London and Stockholm; Nirman Gupta will lead engineering on business agents from London. The stated aim is to help businesses automate operations and run applications reliably as they grow. ### Source excerpt Lovable has acquired the team behind Radish AI, an agentic infrastructure platform. Its co-founders join our engineering team in London and Stockholm to build the agent capabilities and infrastructure behind Lovable. ## Amplifon Americas Data Breach Impacts 1.3 Million DevFeed: [Amplifon Americas Data Breach Impacts 1.3 Million]() Original publisher: [Read original article]() Author: Ionut Arghire Published: 2026-11-09T11:20:09Z Content type: news Language: en Sources: [SecurityWeek]() Topics: [Cybersecurity](), [sensitive data]() Tags: [amplifon](), [breach](), [data-breach](), [data-breaches](), [hackers](), [health](), [healthcare](), [impact](), [information](), [insurance](), [personal](), [social](), [us]() ### AI overview Amplifon Americas disclosed that hackers accessed internal systems in January 2026 and copied files from January 18 through February 8. Its review found personal, government ID, hearing and medical, service, health, and insurance information potentially compromised; some people also had Social Security numbers and financial information stolen. A Texas filing reportedly lists 1,332,826 people affected nationwide. Amplifon says it is unaware of fraudulent use of the information. The company contained the incident and investigated, but took eight months to review the stolen data and determine potential impact. ### Source excerpt Hackers stole personal, health, and insurance information, including Social Security numbers, from the company's internal systems. The post Amplifon Americas Data Breach Impacts 1.3 Million appeared first on SecurityWeek. ## Podcast: Distributed LLM Inference on Kubernetes with llm-d, Wasm, & TornadoVM DevFeed: [Podcast: Distributed LLM Inference on Kubernetes with llm-d, Wasm, & TornadoVM]() Original publisher: [Read original article]() Author: Edoardo Vacchi Published: 2026-11-09T11:00:00Z Content type: news Language: en Sources: [InfoQ]() Topics: [LLM optimization](), [AI Infrastructure](), [vllm](), [Large Language Model](), [Local AI](), [TornadoVM](), [LLM observability](), [Wasm components]() Tags: [ai](), [ai-inference](), [ai-ml-data-engineering](), [artificial-intelligence](), [backend](), [batch](), [cache](), [cloud-computing](), [clusters](), [containers](), [distributed](), [distributed-ai-inference](), [distributed-llm-inference-kubernetes](), [gpu](), [gpu-utilization](), [inference](), [infoq](), [java](), [kubernetes](), [kubernetes-clusters](), [large-language-models](), [llm](), [llm-d](), [llm-inference](), [ml-data-engineering](), [podcast](), [the-infoq-podcast](), [tornadovm](), [wasm](), [web-development](), [webassembly]() ### AI overview llm-d extends vLLM's high-throughput inference from a single machine to distributed deployments, balancing requests across instances while favoring cache locality for conversational KV and prefix caches and avoiding overload. It also provides flow control and request prioritization, and can run on bare metal as well as Kubernetes. The discussion describes batch inference during idle periods to use otherwise underused hardware, and semantic classification to route prompts to right-sized models based on factors such as topic or cost. Wasm is presented as useful for sandboxing and portable workloads, including smaller CPU-suitable models, while larger parallel inference generally benefits from GPUs. TornadoVM translates specially written Java code into GPU kernels; it does not run arbitrary Java frameworks such as Spring on the GPU. ### Source excerpt In this episode, Edoardo Vacchi discusses the evolving landscape of WebAssembly on the backend, running Java workloads on GPUs using TornadoVM, and optimizing distributed AI inference using llm-d and vLLM on Kubernetes. Edoardo explains how llm-d manages cache locality (KV and prefix caches), request routing, and batch inference across Kubernetes clusters to maximize expensive GPU utilization. By Edoardo Vacchi ## Your Router Trusts Its Validator. But Who's Watching the Validator? DevFeed: [Your Router Trusts Its Validator. But Who's Watching the Validator?]() Original publisher: [Read original article]() Author: Ritesh Mukherjee Published: 2026-11-09T10:59:54Z Content type: opinion Language: en Sources: [Cyber Newsroom](), [RIPE Labs]() Topics: [observability](), [Access Control](), [known exploitable vulnerabilities](), [Open Source]() Tags: [academic-research](), [account](), [aspa](), [attack](), [attack-surface](), [authentication](), [bgp](), [bgp-routing](), [bug](), [ca](), [cache](), [crash](), [credentials](), [crl](), [infrastructure](), [open-source](), [policy](), [registries](), [repository](), [router](), [rov](), [rpki](), [tools]() ### AI overview RPKI fail-open behavior protects reachability, but an upstream break can quietly change routes from Valid to NotFound without alerting operators. The article traces the chain from CA account through repository, validator, RTR session, and router policy, citing credential compromise, missing or deleted ROAs, validator crashes, and router policy inconsistencies. It presents RAVEN's `raven rtr monitor` as an independent RTR client that tracks cache payload changes and sync timing against an adaptive baseline, distinguishes full resyncs from incremental updates, and streams route validation changes. The tool can surface unusual VRP movement and route transitions, but it does not yet detect a cache that answers while its serial stops advancing, alert on RPKI-driven changes, compare caches, inspect router policy, or repair problems. It supports TCP and TLS, not RTR over SSH. ### Source excerpt ROV turned your validator into a forwarding policy source. The registries have strengthened their side of the RPKI, but the chain from repository to router runs on your infrastructure. When that chain breaks, it often breaks silently. ## Exploited SonicWall SSL-VPN flaw lets attackers bypass SAML-enforced MFA (CVE-2026-19868) DevFeed: [Exploited SonicWall SSL-VPN flaw lets attackers bypass SAML-enforced MFA (CVE-2026-19868)]() Original publisher: [Read original article]() Author: Zeljka Zorz Published: 2026-11-09T10:30:25Z Content type: news Language: en Sources: [Help Net Security]() Topics: [saml](), [MFA](), [known exploitable vulnerabilities](), [Virtual Private Network](), [credential-based attacks](), [Access Control](), [cve remediation](), [Threat monitoring](), [ransomware]() Tags: [2026](), [about](), [authentication](), [bypass](), [company](), [credentials](), [cve](), [cve-2026-19868](), [don-t-miss](), [exploited](), [factor](), [firewall](), [first](), [fixed](), [hot-stuff](), [lets](), [mfa](), [multi-factor-authentication](), [news](), [saml](), [sonicwall](), [ssl](), [ssl-vpn](), [vpn](), [vulnerability]() ### AI overview CVE-2026-19868 is an exploited SonicOS SSL-VPN authentication enforcement flaw that can let someone with valid alternate-method credentials bypass SAML login controls and any MFA enforced through SAML. It applies only when SAML is enabled for SSL-VPN logins; it is not an unauthenticated bypass. SonicWall fixed it in releases 7.3.4-7013 and higher and 8.2.3-8015, affecting Gen7 and Gen8 appliances on versions 7.3.3-7015 and older or 8.2.2-8015 and older. Gen6 is not listed as affected but is out of support; SonicWall advises upgrading Gen6 or older hardware to Gen8. No workarounds, attack details, or IoCs are provided. Organizations using SAML for VPN logins are advised to compare successful VPN logins with identity-provider sign-in and MFA logs, reset credentials for accounts with unmatched logins, and review what those sessions accessed. CISA added the flaw to its Known Exploited Vulnerabilities catalog; US civilian federal agencies must address it by November 20, 2026. ### Source excerpt SonicWall has fixed an exploited SonicOS SSL-VPN vulnerability (CVE-2026-19868) that may allow attackers with valid credentials to bypass SAML-enforced multi-factor authentication. With no workarounds listed, the company has urged admins to patch quickly. About CVE-2026-19868 CVE-2026-19868 is an authentication enforcement flaw in the SonicOS SSL-VPN service, and is exploitable only if SAML is enabled for SSL-VPN logins. According to SonicWall, a user with valid credentials for an alternate authentication method may be able to bypass ... More -> The post Exploited SonicWall SSL-VPN flaw lets attackers bypass SAML-enforced MFA (CVE-2026-19868) appeared first on Help Net Security. ## How to deliver real data sovereignty DevFeed: [How to deliver real data sovereignty]() Original publisher: [Read original article]() Author: Mary Branscombe Published: 2026-11-09T10:00:00Z Content type: article Language: en Sources: [CIO]() Topics: [digital sovereignty](), [Cloud](), [Confidential Computing](), [Shared Responsibility Model](), [Data Management]() Tags: [aws](), [c-suite](), [c-suite-careers-cio-data-center-data-center-management-data-governance-data-integration-data](), [careers](), [cio](), [cloud](), [cloud-regions](), [confidential-computing](), [control](), [costs](), [customer-data](), [data](), [data-center](), [data-center-management](), [data-governance](), [data-integration](), [data-management](), [data-privacy](), [data-residency](), [data-sovereignty](), [dependencies](), [edition-us](), [encryption](), [encryption-keys](), [eu](), [europe](), [google](), [it-leadership](), [it-management](), [it-strategy](), [language-en](), [microsoft](), [operational](), [oracle](), [privacy](), [security](), [sovereignty](), [storytype-feature]() ### AI overview Data sovereignty requires control over access and encryption keys, not just EU data residency. Cloud providers offer regional storage, audited support workflows, customer-approved access, and confidential computing, but a US parent company may still face legal demands; customer-managed keys and access controls can reduce provider control while adding staffing and operational costs. Private or disconnected infrastructure offers more control but demands greater ongoing security and maintenance resources. Strict staff-location rules can sharply raise costs and constrain connectivity, collaboration, and resilience. The article recommends setting controls according to each workload's risks and requirements, classifying data and mapping where it flows, then balancing sovereignty with operational needs rather than applying maximum isolation everywhere. ### Source excerpt Changing where your data is stored isn't enough to deliver sovereignty. You also need to implement the technologies that enforce real control. Local providers are keen to promise alternatives but US ones are making the effort to give customers more options without losing access to the depth of services they offer. Actually turning the promise of sovereign cloud regions into reality means dealing with encryption keys, access workflows and confidential computing, which also means more costs, dependencies, and operational responsibilities. The obvious concern with using US cloud services is the CLOUD Act, which allows the US government to request access to customer data from a US-owned service, even if it's held in the EU. Those requests are supposed to go through the criminal court system. "When it lands in Europe it headbutts with GPDR and you're involved in legal wrangling as to whether or not that data can be released or not," IDC research director Rahiel Nasir says. He views the US Foreign Intelligence Surveillance Act (FISA) as more problematic because it doesn't go through the normal criminal courts. "That's an executive order," he adds. "They'll just take whatever they need." The increasing number of sovereign public cloud regions promising data residency and local operations can't help with that. Controlling encryption keys Under its EU Data Boundary, Microsoft stores and processes EU customer data and even support logs, at least for core cloud services, inside the EU. AWS has similar commitments for its European Sovereign Cloud, claiming no critical dependencies on infrastructure outside the EU. And like Salesforce, it employs EU-based support staff for its EU operating zone. Google and Oracle have similar dedicated sovereign regions in Europe, and some of the providers plan to offer similar models in other jurisdictions. For admin and support, for example, most tasks are performed through automated, audited workflows, and Microsoft personnel don't have stand ## ChaCha AVX-512 Optimization For Linux 7.5 Yields Up To 99% Improvement On AMD Zen 5 DevFeed: [ChaCha AVX-512 Optimization For Linux 7.5 Yields Up To 99% Improvement On AMD Zen 5]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-11-09T12:35:12Z Content type: news Language: en Sources: [Phoronix]() Topics: [ChaCha](), [Linux Kernel]() Tags: [ai](), [amd](), [avx](), [code](), [desktop-linux](), [halo](), [improvement](), [linux](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [open-source-graphics](), [optimization](), [performance](), [phoronix](), [phoronix-test-suite](), [processors](), [ubuntu-benchmarks](), [ubuntu-hardware](), [zen-5]() ### AI overview A new 16-block AVX-512 ChaCha function in Linux libcrypto can improve bulk encryption performance on recent processors. The existing AVX-512VL path handles up to 8 blocks using 256-bit registers; the patch adds a 512-bit path for inputs larger than 8 blocks, gated by X86_FEATURE_PREFER_YMM. The excerpt reports up to 99% improvement on AMD Zen 5 with the full 512-bit data path and up to 39% on Zen 4 using its 256-bit data path. The change is queued in Eric Biggers' libcrypto-next branch and was expected to be submitted for the Linux 7.5 merge window. ### Source excerpt The Linux kernel's ChaCha symmetric stream cipher code has long had an AVX-512 optimized code path but it was catering to the early generations of Intel CPUs with AVX-512 (Skylake) that suffered from down-clocking under use. But now slated for arrival with Linux 7.5 is tuning the code to perform better on today's AVX-512 Intel and AMD processors that don't suffer from this limitation. In turn the ChaCha performance is as much as 99% faster with current-generation AMD Zen 5 processors... ## AWS Launches TorchNeuron to Run PyTorch Models on AWS Trainium Without Modifications DevFeed: [AWS Launches TorchNeuron to Run PyTorch Models on AWS Trainium Without Modifications]() Original publisher: [Read original article]() Author: Sergio De Simone Published: 2026-11-09T11:00:00Z Content type: release Language: en Sources: [InfoQ]() Topics: [AWS Trainium](), [neuron](), [PyTorch](), [AWS AI chips](), [AI Infrastructure]() Tags: [ai](), [ai-chips](), [ai-infrastructure](), [ai-ml-data-engineering](), [amazon](), [annapurna-labs](), [apis](), [architectures](), [aws](), [aws-torchneuron-trainium-pytorch](), [aws-trainium](), [backend](), [capacity](), [conference](), [developers](), [development](), [distributed-training](), [drug-discovery](), [large-language-models](), [machine-learning](), [ml-data-engineering](), [native](), [news](), [no-code](), [pytorch](), [pytorch-conference](), [torchneuron]() ### AI overview Amazon introduced TorchNeuron, a native PyTorch backend for AWS Trainium that aims to reduce migration work: with PyTorch automatic device detection, the text says switching from CUDA requires replacing `.to('cuda')` with `.to('neuron')`. It supports distributed training APIs including DDP, FSDP2 and DTensor, Torch streams and events, and custom performance-critical operators through the Neuron Kernel Interface. Amazon says it validated 100,000 popular models and more than 800 architectures. The backend is also presented as enabling shared hardware capacity for training and inference. The excerpt does not establish production performance or cost; cited observers say workload-specific performance, cost per token, reproducibility, and total cost of ownership remain to be tested. ### Source excerpt At the recent PyTorch Conference North America 2026, Amazon introduced TorchNeuron, a native PyTorch backend that enables developers to run existing PyTorch code on AWS Trainium AI chips with virtually no code modifications. By Sergio De Simone ## Over 3 Million People Hit by Frontline Education Data Breach DevFeed: [Over 3 Million People Hit by Frontline Education Data Breach]() Original publisher: [Read original article]() Author: Eduard Kovacs Published: 2026-11-09T10:09:02Z Content type: news Language: en Sources: [SecurityWeek]() Topics: [data](), [sensitive data]() Tags: [administration](), [attack](), [central](), [data-breach](), [data-breaches](), [education](), [employees](), [frontline-education](), [government](), [hackers](), [health](), [id](), [ids](), [indiana](), [insurance](), [passport](), [personal-information](), [school](), [software](), [students](), [third-party](), [vulnerability]() ### AI overview Frontline Education reported that its August breach affected 3,085,266 people, more than twice the roughly 1.3 million counted in earlier filings with other states. The company said attackers exploited a vulnerability in an unnamed third-party product, accessed part of its environment, and stole personal information from Frontline Central and School Health Management. Frontline detected the intrusion on August 14 and said it contained the attack the next day. Exposed data varied by person and could include names, government ID numbers, dates of birth, health insurance information, and student IDs. US adults are being offered two years of credit monitoring and identity theft protection; eligible minors receive dark web monitoring. The company has not identified the exploited product or the attackers. ### Source excerpt Frontline Education told Indiana regulators that its August data breach affected over 3 million school district staff and students. The post Over 3 Million People Hit by Frontline Education Data Breach appeared first on SecurityWeek. ## $133 million stolen from French bank via exploited service provider DevFeed: [$133 million stolen from French bank via exploited service provider]() Original publisher: [Read original article]() Author: Sinisa Markovic Published: 2026-11-09T10:03:01Z Content type: news Language: en Sources: [Help Net Security]() Topics: [Security](), [fake account fraud]() Tags: [account](), [attack](), [bank](), [blockchain](), [breach](), [credit-mutuel-arkea](), [cryptocurrency](), [cyberattack](), [data-breach](), [don-t-miss](), [eu](), [exploited](), [france](), [hot-stuff](), [news](), [service]() ### AI overview Crédit Mutuel Arkéa says attackers reached the bank directly through an unnamed service provider's platform and enabled fraudulent transfers. It rejects claims that the breach entered through Fortuneo or customer accounts, describes the attack as sophisticated and contained, and says investigations and recovery efforts are under way. The bank confirmed only the order of magnitude of the reported EUR 118 million ($133 million) loss; some funds have been blocked, but the final amount is not yet known. It temporarily suspended some transactions and warned customers to watch for phishing. Separately, blockchain investigator ZachXBT traced more than $100 million in suspected stolen stablecoins through swaps and exchanges, reporting some conversion to Monero; attribution remains unknown. ### Source excerpt Attackers siphoned an estimated $133 million from Crédit Mutuel Arkéa by using the bank's own account to wire money to other banks, from where it was possibly converted into cryptocurrency. The French cooperative banking group, which has EUR 211.5 billion in total assets and more than 5.8 million customers, confirmed the attack days after blockchain investigator ZachXBT began tracking over $100 million in suspected stolen funds moving out of centralized exchange accounts. The bank disclosed the ... More -> The post $133 million stolen from French bank via exploited service provider appeared first on Help Net Security. ## How to give custom AI agents their own service identity DevFeed: [How to give custom AI agents their own service identity]() Original publisher: [Read original article]() Author: Ritesh Patel Published: 2026-11-09T09:41:36.137587Z Content type: tutorial Language: en Sources: [Nirmata]() Topics: [Authentication](), [least privilege]() Tags: [account](), [agent](), [agents](), [ai](), [ai-agent](), [ai-agents](), [ai-governance](), [api](), [app](), [audit](), [authentication](), [authorization](), [credentials](), [custom](), [github](), [github-apps](), [how-to](), [identity](), [kubernetes](), [rbac](), [secrets-management](), [service](), [tokens]() ### AI overview Give each agent a distinct identity and narrow its grants: use a per-agent Kubernetes service account bound to a narrow Role, or a GitHub App when the agent acts on GitHub. Kubernetes projected tokens can be audience-scoped and rotated; applications must reload them, and services should validate them with TokenReview and check the audience. GitHub installation tokens expire after one hour and can be restricted by repository and permission, but omitted restrictions leave the app's full granted access. Neither identity records which person delegated a task, so keep an audit record linking the agent, task, person, tool call, target, and time. The excerpt notes that its manifests have not been run on a cluster. ### Source excerpt Give each custom AI agent its own service account or GitHub App identity, narrow its grants, use short-lived tokens, and log the person each agent acted for. ## Ethersphere Hijacked With Crypto Stealing Malware Via Github & npm DevFeed: [Ethersphere Hijacked With Crypto Stealing Malware Via Github & npm]() Original publisher: [Read original article]() Author: Moshe Siman Tov Bustan Published: 2026-11-09T09:29:08Z Content type: news Language: en Sources: [OX Security]() Topics: [software supply-chain attack](), [Malware](), [npm](), [GitHub](), [Remote Access Trojan](), [Credential theft](), [npm security]() Tags: [application-security](), [backdoors](), [c2](), [compromise](), [cross-platform](), [crypto](), [crypto-wallet](), [ethereum](), [github](), [key](), [keys](), [malware](), [npm](), [ox-security](), [research](), [software-supply-chain-security](), [supply-chain](), [supply-chain-compromise](), [wallet]() ### AI overview OX Security reports a coordinated compromise of Ethersphere's @ethersphere/bee-js npm package and swarm-desktop GitHub repository, apparently using a stolen developer access key. Malicious code contacts a command server, collects Swarm configuration and encrypted wallet-key data, then installs platform-specific remote access: ScreenConnect on Windows and the evjawbreaker Go RAT on Linux and macOS; unsupported platforms receive a decoy script. The article says the npm version was revoked but the malicious GitHub release and command server were still live at publication. It estimates exposure at 13,378 monthly package downloads, 3,152 through affected direct dependents, 260 through an indirect dependent, and 44 GitHub release downloads. Recommended steps are to check dependencies and lockfiles, assess wallet credentials as potentially compromised if the code ran, and pin verified clean versions. ### Source excerpt Threat actors compromised Ethersphere's npm and GitHub packages to steal crypto wallet keys and deploy cross-platform backdoors Overview OX Security researchers identified a coordinated supply chain compromise affecting the Ethersphere ecosystem, a decentralized storage and communication protocol for Ethereum. Threat actors leveraged a suspected compromised developer access key to inject malicious code into both the [...] The post Ethersphere Hijacked With Crypto Stealing Malware Via Github & npm appeared first on OX Security. ## How to Make a Triangle in CSS (3 Ways) DevFeed: [How to Make a Triangle in CSS (3 Ways)]() Original publisher: [Read original article]() Author: Marko Published: 2026-11-09T09:00:00Z Content type: tutorial Language: en Sources: [Marko Denic Blog]() Topics: [CSS](), [Front end](), [Web Development]() Tags: [arrow](), [border](), [clip-path](), [css](), [gradient](), [speech](), [svg](), [tutorial](), [web-development]() ### AI overview Use `clip-path` as the general-purpose CSS triangle method: define a box and cut it with a three-point `polygon()`. Changing the points makes triangles point in different directions; `aspect-ratio: 1 / 0.866` gives an equilateral triangle. For small, solid-color tooltip or dropdown arrows, the border trick makes a triangle from a zero-sized box with one colored border and adjacent transparent borders; the triangle points away from the colored side. A rotated square can provide an arrow with a real border or shadow, while a hard-stop linear gradient creates a right-triangle background decoration. `clip-path` and the border trick do not retain a triangle border; for that, the excerpt suggests stacking triangles, using a rotated square, or using SVG. For a shadow that follows the visible shape, use `filter: drop-shadow()` rather than `box-shadow`. The excerpt recommends SVG for icons and complex shapes. ### Source excerpt Make triangles in pure CSS with clip-path, the classic border trick, or gradients. Triangles pointing in every direction, equilateral triangles, and tooltip and dropdown arrows. ## Agentic engineering: why I built the AI Fundamentals course DevFeed: [Agentic engineering: why I built the AI Fundamentals course]() Original publisher: [Read original article]() Author: Antoine van der Lee Published: 2026-11-09T08:54:50Z Content type: release Language: en Sources: [SwiftLee]() Topics: [agentic-coding](), [AI-assisted coding](), [Learning](), [Prompt Engineering](), [workflow automation]() Tags: [agentic](), [agentic-coding](), [agentic-engineering](), [agents](), [ai-development](), [ai-fundamentals](), [automation](), [code](), [course](), [software]() ### AI overview Agentic engineering shifts developers from writing most code themselves to describing outcomes, reviewing agent changes, and improving the instructions and checks around agents. The author built AI Fundamentals from daily use across projects and lessons that recurred across tools, combining personal experience with official documentation. It covers planning and review, context and prompts, agent instructions, validation and safety, and scaling work across projects; practice uses a browser based sample website and requires basic development experience, but no specific tool or language. The author reports that, across 1,674 agent sessions, 94.2% of tokens were cache reads and 0.6% were agent replies, and the course teaches learners to analyze their own usage. It has nine modules and 66 lessons, released weekly after the first three; the final module and certificate are scheduled for December 22. The article says the course avoids tool specific settings and fixed prompt recipes in favor of transferable workflow judgment. ### Source excerpt Agentic engineering changed how I build software, and today I'm launching the course I wrote about it: AI Fundamentals. I use agentic engineering and agentic coding for the same shift: you write less code yourself, and spend more time describing outcomes, reviewing changes, and improving the system around your agents. The course has been on ... -> The post Agentic engineering: why I built the AI Fundamentals course appeared first on SwiftLee. ## Linux Seeing Driver Work For MSI Mystic Light On Latest Ryzen Motherboards DevFeed: [Linux Seeing Driver Work For MSI Mystic Light On Latest Ryzen Motherboards]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-11-09T11:20:49Z Content type: news Language: en Sources: [Phoronix]() Topics: [Linux Kernel](), [Linux]() Tags: [chip](), [controller](), [core](), [cpu](), [desktop-linux](), [driver](), [i-o](), [kernel](), [lighting](), [linux](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [misc](), [motherboard](), [msi](), [open](), [open-source](), [open-source-graphics](), [phoronix](), [phoronix-test-suite](), [rails](), [rgb](), [sensors](), [ubuntu-benchmarks](), [ubuntu-hardware](), [usb]() ### AI overview Initial Linux driver patches for MSI Mystic Light focus on exposing voltage readings, rather than RGB controls. On tested motherboards, the USB microcontroller monitors the CPU 1P8, CHIP Core, and VDD MISC rails, which the Super I/O chip does not monitor. Linux RGB control may come in a later patch iteration; the excerpt does not establish that it is implemented. ### Source excerpt MSI's latest Ryzen motherboards with AMD 800 series chipsets feature a new capability marked as Mystic Light for RGB lighting controls of the motherboard as well as compatible graphics cards and cases. MSI Mystic Light control software is only offered for Microsoft Windows but thanks to an open-source developer with the help of Claude Code, a Mystic Light open-source driver is being worked on for the Linux kernel... ## Article: Agentic AI Architecture: Current Challenges and Future Opportunities DevFeed: [Article: Agentic AI Architecture: Current Challenges and Future Opportunities]() Original publisher: [Read original article]() Author: Rafał Gancarz Published: 2026-11-09T10:52:00Z Content type: article Language: en Sources: [InfoQ]() Topics: [AI Architecture](), [Agentic AI](), [software-architecture](), [Multi Agent Systems](), [AI Integration Strategies](), [LLM security](), [observability](), [Cloud Architecture](), [decoupling](), [agent to agent communication](), [A2A protocol](), [Artificial Intelligence]() Tags: [a2a](), [adoption](), [agentic](), [agentic-ai](), [agentic-ai-architecture](), [agentic-ai-architecture-challenges-opportunities](), [agents](), [ai](), [ai-agents](), [ai-architecture](), [ai-coding](), [ai-development](), [architecture](), [architecture-design](), [architectures](), [article](), [artificial-intelligence](), [bounded](), [bounded-context](), [challenges](), [cloud](), [cloud-native](), [code-generation](), [components](), [context](), [context-engineering](), [future](), [large-language-models](), [model-context-protocol-mcp](), [retrieval-augmented-generation]() ### AI overview Agentic AI architecture builds on microservices and cloud-native systems, adding agents that use context, memory, and protocols such as MCP for tools and A2A for agent interoperability. The excerpt highlights bounded contexts to limit agents' access to local data and services, and says coordination strategies must be tested for each task: centralized orchestration performed better for financial reasoning, while decentralized coordination suited web navigation. It identifies risks including prompt injection, poisoned tool outputs, and model-provider data handling; proposed safeguards include capability-scoped permissions and task-scoped credentials. Because agent behavior is nondeterministic, the text recommends structured decision logs, quality, latency and spend metrics, tracing, alerts, and runbooks. Adoption is expected to be gradual, with existing architectures coexisting and platforms absorbing common agent components. ### Source excerpt Agentic AI architecture extends microservices and cloud-native architectures with AI agents that work with context, memory, and protocols such as MCP and A2A. The article covers bounded-context integration, task-dependent coordination strategies, new security risks, and observability needs. It expects gradual adoption alongside existing architectures, with platforms absorbing common components. By Rafał Gancarz ## Admission control vs runtime enforcement for AI agents DevFeed: [Admission control vs runtime enforcement for AI agents]() Original publisher: [Read original article]() Author: Ritesh Patel Published: 2026-11-09T09:41:36.137587Z Content type: article Language: en Sources: [Nirmata]() Topics: [control-plane](), [Access Control]() Tags: [admission](), [agent](), [ai](), [ai-agents](), [ai-governance](), [kubernetes](), [kubernetes-security](), [kyverno](), [policy-as-code](), [runtime](), [runtime-enforcement]() ### AI overview Admission control evaluates Kubernetes API requests after authentication and authorization but before storage, so it can validate or mutate pod objects before they start. It cannot observe a running agent's outbound calls. Runtime enforcement must sit in the call path or process to inspect those actions; gateways or in-process hooks can see tool names and arguments, while kernel or network checks cannot read arguments inside encrypted traffic. Use admission for pod-spec risks, narrow RBAC for Kubernetes API permissions, and runtime checks for model, tool, or data-store calls. Webhook `failurePolicy` controls outage behavior: `Fail` blocks matching requests and is the default; `Ignore` allows them unchecked. The excerpt's Kyverno `NamespacedValidatingPolicy` example requires an `ai-agent` pod label. It recommends testing in `Audit` mode before switching to `Deny`, while noting that the cited Kyverno page does not define those action semantics and they should be confirmed for the deployed release. ### Source excerpt An agent's pods face admission control before they start. Its calls face runtime enforcement after. Here is where each one stops it, and where it doesn't. ## Product showcase: What the Malwarebytes Claude connector reveals about unexpected texts DevFeed: [Product showcase: What the Malwarebytes Claude connector reveals about unexpected texts]() Original publisher: [Read original article]() Author: Anamarija Pogorelec Published: 2026-11-09T09:31:52Z Content type: article Language: en Sources: [Help Net Security]() Topics: [Claude](), [AI connector platforms]() Tags: [agentic-ai](), [anthropic](), [claude](), [cybersecurity](), [link](), [malwarebytes](), [news](), [product-showcase](), [scam](), [scanning](), [test]() ### AI overview The Malwarebytes connector lets Claude check suspicious links, phone numbers, and email addresses against Malwarebytes' threat database, including from uploaded message screenshots. It offers six tools with per-tool Ask, Always allow, or Block permissions, works on free and paid Claude plans, and requires no separate Malwarebytes account. In two screenshot tests with permissions set to Ask, a purported Croatian speeding-fine text returned Unknown for its sender number; Claude identified the Philippine country code and pressure to act as scam signs. Two shortened links in IKEA delivery texts also returned Unknown: one scan showed a redirect toward an IKEA order page, while the other was marked Offline. The redirect offered a detail to verify in the IKEA app without opening the text link. These examples show the connector can provide clues for assessing messages, though the excerpt reports Unknown results for the tested phone number and links. ### Source excerpt The Malwarebytes connector for Claude checks suspicious links, phone numbers, and email addresses without leaving the chat. You can paste a link or upload a screenshot of a message, then ask Claude to check the details against Malwarebytes' threat database. The connector works with free and paid Claude plans and requires no separate Malwarebytes account. Setting permissions for each tool The connector has six tools, including options to look up domain registration details, scan several ... More -> The post Product showcase: What the Malwarebytes Claude connector reveals about unexpected texts appeared first on Help Net Security. ## Google DeepMind's AI Agents Make 1024-Bit RSA Keys Cheaper to Attack DevFeed: [Google DeepMind's AI Agents Make 1024-Bit RSA Keys Cheaper to Attack]() Original publisher: [Read original article]() Author: Eduard Kovacs Published: 2026-11-09T08:35:50Z Content type: news Language: en Sources: [SecurityWeek]() Topics: [Security](), [digital signature](), [agentic workflows]() Tags: [agents](), [ai](), [algorithm](), [attack](), [attacks](), [cost](), [cpu](), [cpu-time](), [cryptographic](), [cryptography](), [data-protection](), [days](), [deepmind](), [digital](), [digital-signatures](), [encryption](), [factor](), [featured](), [google](), [google-deepmind](), [keys](), [rsa](), [rsa-1024]() ### AI overview DeepMind researchers working with UC San Diego and Inria report cheaper attacks on obsolete 1024-bit RSA systems, while saying modern protocols using 2048-bit or larger keys remain secure. For PKCS#1 v1.5 signatures, padding provides no protection when an attacker controls roughly half or more of the signed block. In a sandboxed test, AI agents reduced the CPU time for an earlier forgery attack from 1,380 to 391 core-years. The forgery does not recover the private key, needs about 4 billion signing queries in advance, and fails against devices that hash messages internally. For attackers with less control, such as over a 32-byte hash, the paper estimates a method roughly 1,600 times better than the previous approach. A separate, unimplemented algorithm theoretically makes each additional same-size RSA key about 30 times cheaper to factor after an initial computation costing as much as factoring one key. ### Source excerpt Working with cryptographers, DeepMind's AI agents cut the cost of forging RSA-1024 signatures and found a cheaper way to factor keys. The post Google DeepMind's AI Agents Make 1024-Bit RSA Keys Cheaper to Attack appeared first on SecurityWeek. ## Reason Machines Open-Sources Aspen, a Git Server Built for AI Agents DevFeed: [Reason Machines Open-Sources Aspen, a Git Server Built for AI Agents]() Original publisher: [Read original article]() Author: Asif Razzaq Published: 2026-11-09T08:20:20Z Content type: release Language: en Sources: [MarkTechPost]() Topics: [Git]() Tags: [agent](), [agentic-ai](), [agents](), [ai](), [ai-agent](), [ai-agent-traffic](), [ai-agents](), [ai-infrastructure](), [ai-shorts](), [apache-2-0](), [applications](), [artificial-intelligence](), [aws](), [azure](), [branch](), [cdn](), [ci](), [cluster](), [commit](), [commits](), [compaction](), [editors-pick](), [for-devs](), [git](), [harness](), [language-model](), [machine-learning](), [open-source](), [openai](), [pando](), [python](), [server](), [software-engineering](), [staff](), [tech-news](), [technology]() ### AI overview Aspen is an open-source, self-hosted Git server designed for concurrent AI-agent traffic. It stores branch history in PostgreSQL or SQLite and Git packs in object storage; per-branch compare-and-swap lets pushes to different branches proceed independently, while same-branch races return Git's usual "fetch first" error. It reuses stored packs for fetches and supports plain Git over HTTPS and SSH. Reason Machines reports 567 pushes per second sustained for five minutes on one 8-core server, with 208 ms median and 314 ms p99 latency; these are self-reported results, and the project describes them as indicative. Writes to `main` are much slower, around four per second, and Aspen is pre-alpha, with no releases or web UI. Deployment options described include local development, Docker Compose, Helm, Terraform, and mirror mode for GitHub or GitLab. ### Source excerpt Reason Machines has open-sourced Aspen, a self-hosted Git server for AI agents. It follows OpenAI's Pando design and sustained 567 pushes per second on one 8-core server in the company's own test. The post Reason Machines Open-Sources Aspen, a Git Server Built for AI Agents appeared first on MarkTechPost. ## Worth Reading: BGP Multi-Homed with Two ISPs and Two Routers DevFeed: [Worth Reading: BGP Multi-Homed with Two ISPs and Two Routers]() Original publisher: [Read original article]() Published: 2026-11-09T07:25:00Z Content type: article Language: en Sources: [ipSpace]() Topics: [BGP](), [Network]() Tags: [bgp](), [reading](), [worth-reading]() ### AI overview Suresh Vinasiththamby's BGP multihoming scenario covers practical details for building a working setup with two ISPs and two routers. The recommendation notes that Suresh used netlab to build the lab, making it possible to reproduce the scenario with non-Cisco devices. The excerpt does not specify the configuration details or which BGP requirements the article covers. ### Source excerpt I was reading the nice BGP Multi-Homed with Two ISPs and Two Routers scenario by Suresh Vinasiththamby - Packetswitch and thought wouldn't it be nice if he were using netlab to build the lab, as that would give you the opportunity to build the same lab with zero hassle using non-Cisco devices. Guess what - Suresh already did just that. I was so glad to see someone using netlab in just the way I imagined it to be used. Read more ... ## DeepX Tapes Out Second-Gen Chip, Projects 6.7x Performance Gain With LPDDR5x-PIM DevFeed: [DeepX Tapes Out Second-Gen Chip, Projects 6.7x Performance Gain With LPDDR5x-PIM]() Original publisher: [Read original article]() Author: Sally Ward-Foxton Published: 2026-11-09T06:47:26Z Content type: news Language: en Sources: [EE Times]() Topics: [Chip design](), [Enhancing AI model performance](), [ddr5](), [Intel Core]() Tags: [2-nm](), [accelerators](), [ai](), [ai-accelerator](), [ai-chip](), [ai-chips](), [ai-compute](), [ai-data-center](), [ai-inference](), [ai-pcs](), [ai-processor](), [analog](), [analysis](), [business](), [chip](), [chiplet](), [community](), [designline](), [device](), [digital](), [edge](), [education](), [ee-times](), [ee-times-university](), [eelife](), [eelive-electronic-design](), [eet](), [electrical-engineering](), [electronic-components](), [electronic-engineering-times](), [electronics](), [embedded-systems](), [fundamentals-courses](), [ics](), [learning](), [llms](), [lpddr5x](), [memory](), [mixed-signal](), [news](), [opinion](), [part-search](), [performance](), [pim](), [product-search](), [robotics](), [semiconductors](), [software](), [tech-papers](), [technology](), [times](), [video]() ### AI overview DeepX has taped out its DX-M2 edge AI accelerator, designed to run 20-billion to 100-billion-parameter models at 10-30 tokens per second within a 5 W budget. DeepX attributes gains over the DX-M1 to tripled memory bandwidth (50 to 150 GB/s), NPU improvements, and a move to a 2-nm process. Samsung LPDDR5x-PIM could raise Llama 3.1 8B throughput 6.7x in simulations by offloading GEMV operations, but overall power would rise 2.7x; DeepX says a 2x gain may fit the non-PIM power envelope. PIM hardware testing and compiler work remain in progress, and the article describes the results as projections. DeepX targets robotics, AI PCs, automotive, and IoT. It expects DX-M2 samples in March 2027 and mass production in fall 2027; the planned DX-M3 is slated for 2028 with LPDDR6-PIM. ### Source excerpt DeepX is targeting edge markets including robotics with its second-generation AI accelerator, which can run large LLMs on device in under 5 W. The post DeepX Tapes Out Second-Gen Chip, Projects 6.7x Performance Gain With LPDDR5x-PIM appeared first on EE Times. [Next page]()