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, Friday 16 October 2026 Retrospective Covers 15 October 2026 00:00 to 16 October 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-10-16/ TOP DEVELOPMENTS [1] Is AI pentesting as good as a human pentester? BleepingComputer · Research · https://bleepingcomputer.com/news/security/is-ai-pentesting-as-good-as-a-human-pentester [2] GhostAction Compromises About 1,900 GitHub Accounts, Targets Over 30,000 Repositories The Hacker News · Incident · https://thehackernews.com/2026/10/ghostaction-compromises-about-1900.html [3] AhsayCBS Flaws Exploited to Deploy Web Shells Get New Fixes After Incomplete Patches The Hacker News · Incident · https://thehackernews.com/2026/10/ahsaycbs-flaws-exploited-to-deploy-web.html [4] Next.js Ships Out-of-Band Patch for Two Critical RCE Flaws in Caching and SVG Processing The Hacker News · Advisory · https://thehackernews.com/2026/10/nextjs-ships-out-of-band-patch-for-two.html [5] Drupal Patches Critical Field UI Flaw With Unauthenticated RCE Path via Contributed Modules The Hacker News · Advisory · https://thehackernews.com/2026/10/drupal-patches-critical-core-rce-flaw.html SAST [6] Combining LLM and Dynamic Analysis for Universal Testing of Static Application Security Testing Tools arXiv · Research · https://arxiv.org/abs/2610.21309 SCA [7] Can SOLARBURST go silicon? A systematic analysis on software supply chain attacks for VLSI design flows arXiv · Research · https://arxiv.org/abs/2610.22580 [8] Agentic Dependency Transformations: Lessons from an Exploratory Case Study of Flask arXiv · Research · https://arxiv.org/abs/2610.21372 [9] VulnBump: Can AI Agents Fix Vulnerable Dependencies Without Breaking the Build? Snyk · Research · https://snyk.io/blog/vulnbump AGENTIC SDLC [10] What Survives Self-Evolution: Provenance Marks in Copied Agent Harnesses arXiv · Research · https://arxiv.org/abs/2610.21458 [11] Plausible Is Not Enough: Trust and Evidence in AI-Authored Pull Request Review arXiv · Research · https://arxiv.org/abs/2610.21144 [12] Your AI agent is sandboxed. Its data still needs guardrails. Mend.io · Analysis · https://mend.io/blog/docker-sandboxes-ai-guardrails TRUST [13] EU Cyber Resilience Act: Preparing for December 11, 2027 Endor Labs · Analysis · https://endorlabs.com/learn/the-rest-of-the-eu-cyber-resiliency-act-lands-on-december-11th-2027 [14] MCP Tool Exposure is a Privilege Grant: A Two-Gate Authorization Model for Agentic AI Cloud Security Alliance · Analysis · https://cloudsecurityalliance.org/articles/mcp-tool-exposure-is-a-privilege-grant-a-two-gate-authorization-model-for-agentic-ai [15] Why AI Confidence Is Not Validation The Hacker News · Analysis · https://thehackernews.com/2026/10/why-ai-confidence-is-not-validation.html MODELS [16] Refusal Is Not Robustness: Measuring Harmful Reachability in Autoregressive Language Models arXiv · Research · https://arxiv.org/abs/2610.22754 [17] Shortcut Sensitivity in Agent-Safety Evaluations arXiv · Research · https://arxiv.org/abs/2610.22243 [18] Federated Art Chase: Privacy Leakage and Data Poisoning in Federated Fine-Tuning arXiv · Research · https://arxiv.org/abs/2610.21547 AI SECURITY [19] Jev as a judge: How counterfeit policy steers Jev Datadog · Standard / policy · https://securitylabs.datadoghq.com/articles/jev-as-a-judge-how-counterfeit-policy-steers-jev [20] Proofpoint to buy Varonis for $6.3 billion CyberScoop · Analysis · https://cyberscoop.com/proofpoint-acquires-varonis-systems-data-security [21] My Credentials, Please: Detecting Filesystem Capability Laundering in MCP Servers at Scale arXiv · Research · https://arxiv.org/abs/2610.22301 ADVISORIES IN THIS WINDOW 36 Go advisories · 5 critical · 15 high · 12 moderate · 4 low 17 PyPI advisories · 9 critical · 8 moderate 13 npm advisories · 6 critical · 3 high · 4 moderate 3 Packagist advisories · 2 high · 1 moderate 1 Pub advisory · 1 high 1 crates.io advisory · 1 high 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.