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, Saturday 3 October 2026 Retrospective Covers 2 October 2026 00:00 to 3 October 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-10-03/ TOP DEVELOPMENTS [1] Dell asks admins to patch max severity CSM flaws as soon as possible BleepingComputer · Advisory · https://bleepingcomputer.com/news/security/new-max-severity-dell-csm-flaws-give-hackers-admin-privileges [2] GitLab warns of critical RCE vulnerability in AI Gateway service BleepingComputer · Analysis · https://bleepingcomputer.com/news/security/gitlab-warns-of-critical-rce-vulnerability-in-ai-gateway-service [3] Repository security advisory comments API in public preview GitHub · Advisory · https://github.blog/changelog/2026-10-02-repository-security-advisory-comments-api-in-public-preview SAST [4] Engineering Sustainable Agents: A Systematic Comparison of Agentic LLMs for Developer Workflows arXiv · Research · https://arxiv.org/abs/2610.03010 SCA [5] Post-Quantum Cryptography for Sigstore Sigstore · Analysis · https://blog.sigstore.dev/pqc-sigstore DAST [6] AgentTrap: Stateful Feedback Deception against Autonomous Penetration Testing Agents arXiv · Research · https://arxiv.org/abs/2610.02869 AGENTIC SDLC [7] Beware EviLLM: Enabling Vulnerability Injection via Large Language Models arXiv · Research · https://arxiv.org/abs/2610.03857 [8] Control OSWorld: An AI Control Environment for GUI Computer Use Agents arXiv · Research · https://arxiv.org/abs/2610.03818 REMEDIATION [9] Medical records giant Epic pauses product development to fix security bugs that risk patients’ data TechCrunch: Security · Analysis · https://techcrunch.com/2026/10/02/medical-records-giant-epic-pauses-product-development-to-fix-security-bugs-that-risk-patients-data MODELS [10] Self-Propagating Misalignment in LLM Agents, and Why Auditing or Disabling Memory Is Not Enough arXiv · Research · https://arxiv.org/abs/2610.04083 [11] Auditing the Privacy of Synthetic Gene Expression Data: A Unified Weighted-Distance Framework for No-Box Membership Inference arXiv · Research · https://arxiv.org/abs/2610.04060 [12] Threat-Preserving Representation Sensitivity in Agent-Security Benchmarks arXiv · Research · https://arxiv.org/abs/2610.03585 AI SECURITY [13] CorrectGuard: Eyes-Off Correctness Estimation for Black-Box Security Guardrails arXiv · Research · https://arxiv.org/abs/2610.03470 [14] Persona Guardrail: A Production-Grade Defense Framework for Agentic Systems arXiv · Research · https://arxiv.org/abs/2610.03434 [15] Securing Computer-Use Agents Against Branch Steering Attacks arXiv · Research · https://arxiv.org/abs/2610.03089 ADVISORIES IN THIS WINDOW 53 npm advisories · 40 critical · 9 high · 4 moderate 12 PyPI advisories · 4 critical · 6 high · 2 moderate 9 Go advisories · 3 critical · 3 high · 2 moderate · 1 low 5 crates.io advisories · 2 high · 1 moderate · 2 low 1 RubyGems advisory · 1 moderate 1 Maven advisory · 1 high 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.