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 2 May 2026 Retrospective Covers 1 May 2026 00:00 to 2 May 2026 00:00 UTC Web: https://keryn.dev/headlines/daily/2026-05-02/ TOP DEVELOPMENTS [1] The Productivity-Reliability Paradox: Specification-Driven Governance for AI-Augmented Software Development arXiv · Research · https://arxiv.org/abs/2605.01160 [2] A Low-Latency Fraud Detection Layer for Detecting Adversarial Interaction Patterns in LLM-Powered Agents arXiv · Research · https://arxiv.org/abs/2605.01143 [3] LLM Ghostbusters: Surgical Package Hallucination Suppression via Adaptive Unlearning arXiv · Research · https://arxiv.org/abs/2605.01047 SCA [4] Poisoned Ruby Gems and Go Modules Exploit CI Pipelines for Credential Theft The Hacker News · Analysis · https://thehackernews.com/2026/05/poisoned-ruby-gems-and-go-modules.html DAST [5] ClozeMaster: Fuzzing Rust Compiler by Harnessing LLMs for Infilling Masked Real Programs arXiv · Research · https://arxiv.org/abs/2605.00413 [6] ZAP Updates - April 2026 ZAP (Checkmarx) · Release · https://zaproxy.org/blog/2026-05-01-zap-updates-april-2026 TRUST [7] Semia: Auditing Agent Skills via Constraint-Guided Representation Synthesis arXiv · Research · https://arxiv.org/abs/2605.00314 MODELS [8] SRTJ: Self-Evolving Rule-Driven Training-Free LLM Jailbreaking arXiv · Research · https://arxiv.org/abs/2605.00974 [9] RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs arXiv · Research · https://arxiv.org/abs/2605.02946 [10] Jailbreaking Vision-Language Models Through the Visual Modality arXiv · Research · https://arxiv.org/abs/2605.00583 AI SECURITY [11] Self-Adaptive Multi-Agent LLM-Based Security Pattern Selection for IoT Systems arXiv · Research · https://arxiv.org/abs/2605.00741 [12] CleanBase: Detecting Malicious Documents in RAG Knowledge Databases arXiv · Research · https://arxiv.org/abs/2605.00460 ADVISORIES IN THIS WINDOW 6 npm advisories · 5 critical · 1 moderate 3 PyPI advisories · 2 high · 1 moderate 3 Maven advisories · 2 high · 1 moderate 2 Packagist advisories · 2 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.