the xtralinux.org desk
The open web. On watch.
Brief, source-linked news about security, linux and open source — tracked and explained as it breaks, by the XtraLinux desk.
Latest entriesGoogle GTIG Reports Agentic AI Threat Activity Crossing a Significant Threshold
GTIG's latest AI Threat Tracker says threat actors have moved to autonomous multi-agent frameworks like ARTEX and PentAGI, fully automated LLM exploit pipelines, and LLMJacking monetization through underground LLM Mixers.
Amazon MadPot Decoys Capture AI Agents Running API-Key Theft and Deepfake KYC Bypass Campaigns
Amazon captured AI agents' records from campaigns assessed with moderate confidence as running on ARTEX: one stole provider API keys from exposed AI apps to fund more hunting, another used deepfakes to beat exchange KYC checks.
Midnight Blue Finds Six Unpatched Pre-Auth RCE Flaws in Danfoss AK-SM800A Refrigeration Controllers
Six unpatched pre-auth RCE flaws in Danfoss AK-SM800A controllers let an attacker take over the device as root; no fix exists and Belgian, Irish and German agencies are telling operators to stop exposing the controllers online.
Ubuntu 27.04 "Tenacious T-Rex" Daily Builds Open the Development Cycle
First Ubuntu 27.04 daily build ISOs are on cdimage.ubuntu.com, seeded from 26.10 with Linux 7.3 and GNOME 51; GNOME 52 and a newer kernel are due by the April 22, 2027 final release.
MongoDB CVE-2026-118383: Critical Pre-Auth Use-After-Free Crashes Lagging Replica Set Primaries
CVE-2026-118383 (CVSS v4.0 9.5): two unauthenticated ping requests can set up a use-after-free that crashes a MongoDB replica set member if it later wins an election while lagging; fixed in 9.0.7, 8.3.16, 8.0.37 and more.
Cursor Traces Trojanized Claude Code Extension to Stolen Fireworks AI Access Key
Cursor's investigation update: an attacker used an access key exposed in the Fireworks AI breach to reach Cursor data and publish trojanized Anthropic.claude-code 2.1.312/2.1.313 builds; incident contained as of November 1.
Mior and Skyfall Break AMD SEV-SNP Confidential VMs: Hypervisor-Level Memory Read/Write and Fake Attestation
Two new AMD SEV-SNP attacks — Mior (CVE-2026-0462) and Skyfall (CVE-2026-40671) — let a malicious host read or modify confidential VM memory and forge attestation. Fixes ship via SEV firmware and platform BIOS updates.
PostgreSQL November Update Fixes 52 Flaws Including 22-Year-Old Type Confusion RCE (CVE-2026-6460)
PostgreSQL 18.7/17.12/16.16/15.20/14.25 fix 52 vulnerabilities, led by type-confusion RCE CVE-2026-6460 (CVSS 8.8). C extensions must be rebuilt; PG 14 goes end of life.
UnrealIRCd Server-Level Command Injection Flaw (CVE-2026-114713) Exploited in the Wild, Added to CISA KEV
CISA added CVE-2026-114713 (CVSS 8.6), a labeled-response module flaw letting any user inject server-to-server commands on linked UnrealIRCd 6.0.0–6.2.7.0 networks, to KEV after active exploitation.
Citrix Patches 12 NetScaler and Secure Access Client Flaws, Including Console Authentication Bypass and Root Privilege Escalation
Citrix fixed 12 flaws across NetScaler ADC, Gateway, Console and Secure Access Client: memory-overflow DoS bugs up to CVSS 9.5, a 9.4 Console authentication bypass, and root escalation in the Linux client. No exploits known.
Check Point Gateways Under Active Attack: PHP2Shell Plus Gaia Privilege Escalation (CVE-2026-123039) Lead to TSGA Backdoor and ESXi Wiper
Unknown attacker chains PHP2Shell and Gaia privesc CVE-2026-123039 on unpatched Check Point gateways since October 30, deploying TSGA backdoor, webshells and an ESXi wiper. Fixed via JHF takes; EOS branches get a mitigation only.
Red Hat Quay CVE-2026-123710: Forged Signing Keys Give Attackers Registry-Wide Access to Private Container Images
Red Hat rates CVE-2026-123710 in Quay critical at 9.8: an anonymous token tricks key rotation into trusting an attacker key with no expiry, so forged tokens expose and overwrite private images. A 9.9 flaw leaks robot tokens.