The Wire
Curated cybersecurity and tech news — AI-summarized, source attributed.
Glasswing Locks Down the Code, But Your Stack's Exposure Is Still Yours to Own
Glasswing's pitch centers on hardening application code itself, closing off a class of vulnerabilities at the source layer. That narrows one attack surface, but
Hidden Bluetooth tracker in postcard exposes Dutch warship's location
A journalist at Dutch outlet Omroep Gelderland followed a public mailing address on the Dutch government website and slipped a Bluetooth tracker inside a postca
Latin America Overtakes Africa as Most-Attacked Region in Q1 2026
Regional cyberattack rankings shifted this quarter, with Latin America surpassing Africa as the most-targeted region globally. The change reflects a measurable
Lazarus Group Pivots ClickFix Social Engineering to macOS Targets
North Korea's Lazarus Group has extended its ClickFix campaign to macOS, expanding a technique previously aimed at Windows users. ClickFix relies on social engi
Microsoft Entra passkeys land on Windows, closing a gap on unmanaged devices
Microsoft is rolling out Entra passkey support on Windows starting late April, with general availability targeted for mid-June 2026. The feature lets users crea
Pack2TheRoot: 12-year-old PackageKit flaw hands local users root on most Linux distros
A high-severity vulnerability in PackageKit, the daemon that brokers software install and update requests across most Linux distributions, lets unprivileged loc
Project Glasswing: AI Finds the Bugs, But Humans Still Bottleneck the Fix
Project Glasswing demonstrated that AI systems can surface software vulnerabilities at a pace and scale human auditors cannot match. The finding pipeline is no
Rituals breach exposes member data from 41M-strong loyalty database
Dutch cosmetics retailer Rituals confirmed that an attacker exfiltrated personal records from its My Rituals loyalty program database earlier this month. Expose
The AI Agent Authority Gap: Why Continuous Observability Must Drive Runtime Decisions
Enterprise AI agents are being granted production-level permissions—executing trades, modifying records, calling internal APIs—without the runtime controls that
ThreatsDay Roundup: $290M DeFi Heist, macOS LotL Abuse, ProxySmart SIM Farms
The latest ThreatsDay bulletin spans a wide threat surface in a single news cycle. A $290M DeFi exploit dominates the financial-crime beat, underscoring that on
Trigona ransomware deploys custom exfiltration tool to dodge detection
Trigona ransomware operators have shifted from off-the-shelf utilities like Rclone and MegaSync to a bespoke command-line tool called uploader_client.exe, accor
Tropic Trooper Weaponizes SumatraPDF and GitHub to Drop AdaptixC2
The Tropic Trooper APT group is abusing a trojanized build of the open-source SumatraPDF reader as a delivery vehicle for AdaptixC2, a newer command-and-control