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, Monday 27 October 2025 Retrospective Covers 26 October 2025 00:00 to 27 October 2025 00:00 UTC Web: https://keryn.dev/headlines/daily/2025-10-27/ TOP DEVELOPMENTS [1] Scalable Supervision for Software Agents via Patch Reasoning arXiv · Research · https://arxiv.org/abs/2510.22775 [2] Sentra-Guard: A Real-Time Multilingual Defense Against Adversarial LLM Prompts arXiv · Research · https://arxiv.org/abs/2510.22628 [3] Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents arXiv · Research · https://arxiv.org/abs/2510.22620 AGENTIC SDLC [4] The Productivity–Security Paradox of AI Coding Assistants Checkmarx · Analysis · https://checkmarx.com/blog/ai-llm-tools-in-application-security/the-productivity-security-paradox-of-ai-coding-assistants MODELS [5] Cross-Paradigm Graph Backdoor Attacks with Promptable Subgraph Triggers arXiv · Research · https://arxiv.org/abs/2510.22555 ADVISORIES IN THIS WINDOW 1 PyPI 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.