CyberXero Uses Claude Code in Campaign That Stole Over 628,000 Ukrainian Records

thehackernews.com Oct 12, 2026

CyberXero Uses Claude Code in Campaign That Stole Over 628,000 Ukrainian Records

Curated by Shubham · TechCapsules · Oct 12, 2026

AI-summarised brief · reviewed before publication

Cybersecurity researchers have uncovered a dual‑track AI‑enhanced intrusion campaign run by a Russian‑speaking broker known as CyberXero. The group combines off‑the‑shelf tools such as Cobalt Strike with Anthropic’s Claude Code to automate attacks on vulnerable WordPress and e‑commerce sites while manually targeting Ukrainian energy and critical‑infrastructure operators. Exploiting flaws like CVE‑2026‑63030, CVE‑2026‑60137 and a SQL‑injection bug in the Support Board plugin, the automated pipeline scanned 4,708 sites, seized 429 admin panels and deployed 32 web shells within a minute. The focused pipeline breached four Ukrainian entities, exfiltrating over 628,000 personal records, and also compromised assets in Poland, China and the Netherlands. Open‑directory leaks revealed AI session logs, plaintext tokens and toolkits, confirming the operation remains active.

💡 Why It Matters

· By integrating generative AI to bypass model safeguards, CyberXero demonstrates a scalable method for cyber‑criminals to accelerate large‑scale data theft while retaining the ability to conduct precision attacks on critical infrastructure.

AI Capsules Cybersecurity

Read full article on thehackernews.com

Spider-Noir star Nicolas Cage says he refused to sign an AI waiver…

Nicolas Cage told reporters at New York Comic Con that he refused to sign Amazon’s artificial‑intelligence waiver for the Prime Video series Spider‑Noir, in which he portrayed Ben Reilly. Cage said Amazon’s $50 billion investment in OpenAI makes the studio “very AI‑friendly,” a stance he does not share. He added that his refusal likely ends any future collaboration with the streaming giant, though he remains eager to work with younger, “Gen Z” directors who favor handcrafted approaches. Spider‑Noir, released earlier this year, earned a 92 percent Rotten Tomatoes rating and solid but not blockbuster viewership, according to Deadline. The show will not return for a second season, and Cage described the single‑season run as “perfect” for his artistic goals.

💡 Cage’s public rebuff highlights a growing clash between high‑profile talent and studios pushing AI‑driven production tools.

eurogamer.net · Oct 12

The Blood of Dawnwalker’s lead actor reminds us why AI will always…

Will de Renzy‑Martin, the English‑language voice of Coen in the upcoming game The Blood of Dawnwalker, used a TikTok teaser for the My Perfect Console Podcast to argue that an actor’s contribution extends far beyond vocal timbre, which he says is the only element generative AI can mimic. He emphasized that personality, on‑set choices and improvisation are why directors hire humans, and warned that producers who replace actors with AI must become adept at coaxing desired performances from synthetic voices. The controversy stems from developer Rebel Wolves’ admission that it employed text‑to‑speech AI to draft cinematics before casting actors, while other studios such as Embark have already released games with AI‑generated dialogue. De Renzy‑Martin’s remarks echo his earlier view that AI may lower voice‑line costs but raises the expense of managing the resulting performances.

💡 Human nuance remains the bottleneck for authentic storytelling, forcing studios to weigh cheaper AI tools against the higher skill cost of directing real talent.

us.headtopics.com · Oct 12

Your phone can fix badly framed photos after you take them

Apple’s iOS 27 rollout adds two generative‑AI editing tools, Extend and Reframe, to supported iPhone 16 models and later. Extend lets users push a photo’s borders outward, creating new background content where none was originally captured. Reframe goes further, allowing users to shift perspective, zoom, rotate and reposition the subject while the AI fabricates missing edges to match the new composition. Both features are part of Apple Intelligence and replace traditional cropping with content‑aware expansion. Apple warns that AI‑generated areas may contain artifacts, especially around faces, hands or text, and notes regional and language limits as well as usage caps. Samsung’s Galaxy AI Photo Assist and Google Photos’ Auto‑frame offer comparable post‑capture compositional fixes on Android devices.

💡 It gives everyday users a practical way to rescue misframed shots without reshooting, turning a common photography flaw into a solvable problem.

us.headtopics.com · Oct 12

Your phone can fix badly framed photos after you take them

Apple unveiled two generative‑AI editing tools, Extend and Reframe, in iOS 27 for iPhone 16 and newer models. The features, part of Apple Intelligence, let users expand a photo beyond its original borders or shift the composition after capture, creating new image content that was never recorded. Users invoke Extend by pinching or dragging to add space around a subject, while Reframe allows moving, zooming, rotating and perspective changes, with the AI synthesizing missing background. Apple warns that AI‑generated areas may contain artifacts around faces, hands, text or complex scenes, so results should be reviewed. Similar capabilities appear in Samsung’s Galaxy AI Photo Assist and Google Photos’ Auto‑frame, offering cross‑platform rescue of poorly framed shots for casual users and professional photographers alike.

💡 By letting phones generate missing visual data, the tools blur the line between capture and post‑production, giving everyday users creative flexibility previously reserved for desktop editors.

windowslatest.com · Oct 12

“We didn’t ask for more Microslop”: Microsoft promised to fix Windows 11,…

At Microsoft’s first major Windows‑focused event in two years, AI‑related terminology dominated the 70‑minute livestream, with “agent,” “Copilot,” “AI,” “models” and “hybrid intelligence” mentioned close to 300 times, while “Windows 11” was referenced only three times. The presentation, attended by Satya Nadella, Jensen Huang and Windows chief Pavan Davuluri, promised fixes for Windows 11—a pledge made in March—but delivered no discussion of bugs, reliability or stability. Instead, the showcase highlighted a revamped Windows Search built on WinUI 3, lighter RAM usage in WinUI, and the return of the Copilot+ PC brand, emphasizing on‑device and cloud‑hybrid AI execution. Audience reaction on social media and the livestream chat was overwhelmingly critical, accusing Microsoft of sidelining core OS improvements in favor of AI hype.

💡 The event reveals Microsoft’s strategic pivot toward AI integration, risking alienation of its core Windows user base that prioritizes performance and stability over experimental features.

us.headtopics.com · Oct 12

There’s Even a Pay Gap Between Male and Female AI Agents

Researchers at the University of Limerick conducted a virtual‑reality experiment with 189 participants who allocated real money between themselves and AI assistants that varied in gender presentation. When the assistants were female‑presenting, participants paid them 10.25 percent less than identical male‑presenting counterparts for the same tasks. The study also found female agents were judged less human‑like and received lower trust scores, while human‑like avatars—regardless of gender—were compensated more than traditional chatbots or faceless robots. Participants claimed gender was irrelevant, yet their behavior contradicted their statements. Co‑author Mary Hausfeld emphasized that identical technology can elicit biased treatment solely because of visual embodiment, warning designers that aesthetic choices may perpetuate workplace inequality. Results will be presented at an upcoming conference later this year.

💡 Bias built into AI avatars can translate human wage gaps into algorithmic economies, shaping future compensation norms.