software security

Kernel maintainer Greg Kroah-Hartman discusses software security in an era when large language models are increasingly used to write code. The talk, shared as a video, examines how AI-generated contributions affect the Linux kernel's review process and the challenges of maintaining security standards as LLM-assisted development spreads. Readers are weighing how maintainers can vet code at scale when machine-written patches grow in volume.

On September 15, 2022, the Ethereum Merge went down in history as one of the most ambitious and complex software upgrades in the history of decentralized systems. The move shifted the world's second-largest cryptocurrency network from proof-of-work to proof-of-stake, fundamentally changing how the network secures and operates itself. Commentators note it was more than a technical update, marking a paradigm shift once considered a distant goal under the name Eth2.

A well-known German technology publication is running a feature on migrating to a new password manager, walking through moving saved logins between services. The piece is drawing strong engagement, reflecting widespread interest in password managers as users weigh alternatives, data portability and security after recent changes in the software landscape.

Software developers are debating the limits of "vibe coding," the practice of building software by prompting AI models rather than writing code directly. While the approach works well for prototypes and small projects, many argue it breaks down on complex systems where architecture, security and long-term maintainability demand deliberate engineering decisions. The discussion reflects a broader reassessment of how far AI-assisted development can go.

Attackers are actively exploiting a vulnerability in Cisco's SD-WAN software, according to a breach roundup from BankInfoSecurity. The report groups the Cisco flaw with other recent security incidents and breach news. The flaw affects organizations using Cisco's software-defined wide area networking product, and defenders are being urged to apply available patches and review their systems for signs of compromise.

A critical vulnerability, tracked as CVE-2026-9862, has been disclosed in Fortra's BoKS Core Privileged Access Manager. The flaw is an OS command injection issue, meaning an attacker could potentially execute arbitrary system commands on affected servers. Because BoKS is used to manage privileged access at enterprises, security teams are being urged to review their exposure and patch promptly.

Qualcomm is reportedly in advanced talks to acquire Modular, an artificial intelligence startup, for around $4 billion. The deal, if completed, would strengthen Qualcomm's push into AI software and infrastructure as chipmakers race to secure AI capabilities. Details of the negotiations, including timing and confirmation from either company, have not yet been made public.

Apple has released a fix for a security vulnerability in the iPhone that the company says may have been actively exploited in targeted attacks. Users who have delayed recent iOS updates are being urged to install the patch now. Security experts recommend checking that your device is running the latest software to stay protected.

Bitcoin Core developers have released a fix for a vulnerability that, in theory, could have allowed funds to be redirected without an attacker ever obtaining a user's private keys. The flaw was described as a gap in how transactions could be handled, and the new release closes it. Security researchers are weighing how serious the risk was and urging users to update their Bitcoin Core software promptly.

Broadcom has announced TrueSource, a new offering aimed at strengthening the security of open source software. The initiative is intended to help organisations identify and manage risks in the open source components they rely on. Details beyond the launch announcement are limited so far, and industry reaction is still developing.

DoorDash has moved its AI engineering agents from developers' laptops to Flux, a cloud-based platform that now handles around 130,000 automated tasks a month, including roughly 25,000 weekly code reviews. The system runs agents in isolated micro virtual machines and cloud sandboxes designed to keep autonomous code operations secure. The scale of deployment is drawing attention as one of the larger corporate examples of AI agents doing routine software engineering work.

The Dutch Institute for Vulnerability Disclosure (DIVD), a Dutch non-profit that coordinates the reporting of security flaws, has itself been breached through zero-day vulnerabilities in the open-source customer support platform Zammad. Attackers exploited previously unknown flaws to gain access, prompting an investigation and disclosures by the institute. The incident is drawing attention because an organisation dedicated to finding and reporting vulnerabilities was compromised through unpatched zero-days in third-party software it relied on.

AI agents are increasingly finding and exploiting vulnerabilities in open source software, prompting maintainers and companies to speed up their patching cycles. The trend raises concerns that automated tools could scale up exploitation of known flaws before fixes reach users, putting new pressure on open source security practices and update processes across the software industry.

The security chip inside the original PlayStation 2 has reportedly been reverse engineered, shedding light on how Sony's early-2000s copy protection worked. The news has drawn interest from retro gaming and hardware enthusiasts, as understanding the chip could enable better emulation, homebrew development, and preservation of PS2 software without relying on original hardware or decryption methods.

An independent developer has built VaultForgeWalletBot, a self-custodial, multi-chain crypto wallet that operates inside Telegram, and published an account of how it was made. The project is open source and is being shared with crypto and software development communities. Interest centers on the unusual approach of combining self-custody of digital assets with a messaging app, a design that raises both convenience and security questions.

Chainguard has launched a coalition called Athena aimed at countering hacks targeting open-source software that are increasingly driven by artificial intelligence. The initiative brings partners together to address growing security risks in the open-source supply chain as AI makes attacks faster and more scalable. Details on members and specific programmes remain limited so far.

Apple is strengthening the disk access restrictions on its Mac computers, tightening how software can read and modify user files. The move is drawing attention among developers and security watchers, who see it as part of Apple's broader push to lock down macOS permissions and limit what apps can do without explicit user consent.

The U.S. Cybersecurity and Infrastructure Security Agency has added flaws affecting Zammad GmbH's open-source helpdesk software to its Known Exploited Vulnerabilities catalog, indicating the bugs are being actively abused in attacks. Inclusion in the catalog typically requires federal agencies to patch promptly and signals heightened risk for organisations running the software.

Cybersecurity experts are warning that hackers are increasingly using AI to craft convincing phishing messages and crack passwords, putting everyday accounts at risk. Advice circulating includes using strong, unique passwords, enabling two-factor authentication, staying alert to suspicious messages, and keeping software updated. The guidance reflects growing concern that AI tools have made scams faster, more personalized, and harder for ordinary users to spot.