PREVIEW. This is how it will look in an inbox; it was not sent to you. Each real email carries its own unsubscribe link and Keryn's postal address. KERYN HEADLINES Daily, Wednesday 4 November 2026 Retrospective Covers 3 November 2026 00:00 to 4 November 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-11-04/ TOP DEVELOPMENTS [1] Google: LLMjacking attacks rack up $1 million in cloud AI charges BleepingComputer · Incident · https://bleepingcomputer.com/news/security/google-llmjacking-attacks-rack-up-1-million-in-cloud-ai-charges [2] Cyber chief: UK facing AI-powered attacks on critical infrastructure within two years UK National Cyber Security Centre · Incident · https://ncsc.gov.uk/news/cyber-chief-uk-facing-ai-powered-attacks-on-critical-infrastructure-in-two-years [3] Red Hat Patches 9.8-Rated Linux TCP RCE Flaw While Ubuntu Releases Fixes for Testing The Hacker News · Advisory · https://thehackernews.com/2026/11/red-hat-patches-98-rated-linux-tcp-rce.html [4] Synology NAS Devices Targeted in Ransomware Attacks Exploiting Patched DSM Flaw The Hacker News · Advisory · https://thehackernews.com/2026/11/synology-nas-devices-targeted-in.html SCA [5] Hardening GitHub Actions Against Outside Threats Socket · Analysis · https://socket.dev/blog/hardening-github-actions-against-outside-threats [6] Composer 2.11 Adds Dependency Cooldowns Socket · Release · https://socket.dev/blog/composer-dependency-cooldowns [7] Binance Reveals How Fake Job Interviews and AI Tool Abuse Led to $6.43 Billion Theft The Hacker News · Analysis · https://thehackernews.com/2026/11/binance-reveals-how-fake-job-interviews.html AGENTIC SDLC [8] DoGNAVY-Exploit: Harnessing Parallel Open-Weight Agents for Stage-Aware Vulnerability Exploitation arXiv · Research · https://arxiv.org/abs/2611.04196 [9] AI supply chain attacks: Old payloads, new entry points Mend.io · Analysis · https://mend.io/blog/ai-supply-chain-attacks-new-entry-points REMEDIATION [10] Barracuda warns of critical flaws allowing root access to devices BleepingComputer · Analysis · https://bleepingcomputer.com/news/security/barracuda-warns-of-critical-flaws-allowing-root-access-to-devices [11] The PHP Security Wake-Up Call: Fixes Are Only the Beginning OpenSSF · Analysis · https://openssf.org/blog/2026/11/03/the-php-security-wake-up-call-fixes-are-only-the-beginning [12] UEFI bootkit infected fully patched Windows machines. Is Microsoft to blame? Ars Technica: Security · Analysis · https://arstechnica.com/security/2026/11/uefi-bootkit-infected-fully-patched-windows-machines-is-microsoft-to-blame TRUST [13] Whose Transparency is it Anyway? Understanding Generative AI Transparency Needs and Gaps Across Stakeholders arXiv · Research · https://arxiv.org/abs/2611.05549 [14] Stop Chasing CVEs: Start Governing Software Dependencies Sonatype · Analysis · https://sonatype.com/blog/stop-chasing-cves-start-governing-software-dependencies MODELS [15] Extracting System Prompt of LLM Services in Complex (Agentic) Settings: Hierarchical Optimization and Reliable Evaluation Metrics arXiv · Research · https://arxiv.org/abs/2611.04803 [16] BrainwashedLLM: Jailbreaking via Automated Adversarial Chain of Thoughts Discovery arXiv · Research · https://arxiv.org/abs/2611.04717 AI SECURITY [17] Gartner Predicts SOCs Will Have 50% More Engineers Than Analysts by 2028 BleepingComputer · Analysis · https://bleepingcomputer.com/news/security/gartner-predicts-socs-will-have-50-percent-more-engineers-than-analysts-by-2028 [18] ScamWorld: An Interactive Environment for Evaluating Web Agents Against Online Fraud arXiv · Research · https://arxiv.org/abs/2611.04879 [19] Right Tool, Wrong Parameters: Defending LLM Agents Against Parameter Poisoning arXiv · Research · https://arxiv.org/abs/2611.04673 ADVISORIES IN THIS WINDOW 110 npm advisories · 109 critical · 1 high 4 PyPI advisories · 3 critical · 1 high 3 Go advisories · 1 high · 1 moderate · 1 low 1 Maven advisory · 1 high 1 NuGet advisory · 1 high 1 RubyGems advisory · 1 critical All advisories: https://keryn.dev/headlines/advisories/ Titles, links and short excerpts only; every item links to its source. Preview: the unsubscribe link and Keryn's postal address are added to each email when it is sent.