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 13 May 2026 Retrospective Covers 12 May 2026 00:00 to 13 May 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-05-13/ TOP DEVELOPMENTS [1] Mini Shai-Hulud Strikes Again: TanStack + more npm Packages Compromised Wiz · Incident · https://wiz.io/blog/mini-shai-hulud-strikes-again-tanstack-more-npm-packages-compromised [2] Microsoft May 2026 Patch Tuesday fixes 120 flaws, no zero-days BleepingComputer · Advisory · https://bleepingcomputer.com/news/microsoft/microsoft-may-2026-patch-tuesday-fixes-120-flaws-no-zero-days [3] Undermining the trust boundary: Investigating a stealthy intrusion through third-party compromise Microsoft · Incident · https://microsoft.com/en-us/security/blog/2026/05/12/undermining-the-trust-boundary-investigating-a-stealthy-intrusion-through-third-party-compromise SAST [4] Safety Context Injection: Inference-Time Safety Alignment via Static Filtering and Agentic Analysis arXiv · Research · https://arxiv.org/abs/2605.11664 SCA [5] FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems arXiv · Research · https://arxiv.org/abs/2605.11514 [6] Under the Hood of SKILL.md: Semantic Supply-chain Attacks on AI Agent Skill Registry arXiv · Research · https://arxiv.org/abs/2605.11418 [7] RubyGems Suspends New Signups After Hundreds of Malicious Packages Are Uploaded The Hacker News · Analysis · https://thehackernews.com/2026/05/rubygems-suspends-new-signups-after.html DAST [8] Cochise: A Reference Harness for Autonomous Penetration Testing arXiv · Research · https://arxiv.org/abs/2605.11671 AGENTIC SDLC [9] Do Androids Dream of Breaking the Game? Systematically Auditing AI Agent Benchmarks with BenchJack arXiv · Research · https://arxiv.org/abs/2605.12673 [10] Elastic Security MCP App: Interactive security operations inside your AI Tools Elastic · Analysis · https://elastic.co/security-labs/blog/elastic-security-mcp-app [11] OpenAI Launches Daybreak for AI-Powered Vulnerability Detection and Patch Validation The Hacker News · Announcement · https://thehackernews.com/2026/05/openai-launches-daybreak-for-ai-powered.html REMEDIATION [12] Microsoft releases Windows 10 KB5087544 extended security update BleepingComputer · Analysis · https://bleepingcomputer.com/news/microsoft/microsoft-releases-windows-10-kb5087544-extended-security-update [13] Fortinet warns of critical RCE flaws in FortiSandbox and FortiAuthenticator BleepingComputer · Analysis · https://bleepingcomputer.com/news/security/fortinet-warns-of-critical-rce-flaws-in-fortisandbox-and-fortiauthenticator [14] Windows 11 KB5089549 & KB5087420 cumulative updates released BleepingComputer · Analysis · https://bleepingcomputer.com/news/microsoft/windows-11-kb5089549-and-kb5087420-cumulative-updates-released TRUST [15] Attacks and Mitigations for Distributed Governance of Agentic AI under Byzantine Adversaries arXiv · Research · https://arxiv.org/abs/2605.12364 [16] Digital Identity for Agentic Systems: Toward a Portable Authorization Standard for Autonomous Agents arXiv · Research · https://arxiv.org/abs/2605.11487 MODELS [17] GraphIP-Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It? arXiv · Research · https://arxiv.org/abs/2605.12827 [18] REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations arXiv · Research · https://arxiv.org/abs/2605.12813 [19] SoK: A Comprehensive Analysis of the Current Status of Neural Tangent Generalization Attacks with Research Directions arXiv · Research · https://arxiv.org/abs/2605.12792 AI SECURITY [20] Defense at AI speed: Microsoft’s new multi-model agentic security system tops leading industry benchmark Microsoft · Analysis · https://microsoft.com/en-us/security/blog/2026/05/12/defense-at-ai-speed-microsofts-new-multi-model-agentic-security-system-tops-leading-industry-benchmark [21] The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems arXiv · Research · https://arxiv.org/abs/2605.22842 [22] IPI-proxy: An Intercepting Proxy for Red-Teaming Web-Browsing AI Agents Against Indirect Prompt Injection arXiv · Research · https://arxiv.org/abs/2605.11868 ADVISORIES IN THIS WINDOW 151 npm advisories · 138 critical · 7 high · 6 moderate 17 PyPI advisories · 8 critical · 7 high · 2 moderate 10 Go advisories · 1 critical · 6 high · 2 moderate · 1 low 10 Maven advisories · 4 critical · 5 high · 1 low 2 Multi-ecosystem advisories · 2 critical 1 Hex advisory · 1 moderate 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.