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, Tuesday 16 September 2025 Retrospective Covers 15 September 2025 00:00 to 16 September 2025 00:00 UTC Web: https://keryn.dev/headlines/daily/2025-09-16/ TOP DEVELOPMENTS [1] When ‘minimal impact’ isn’t reassuring: lessons from the largest npm supply chain compromise CyberScoop · Incident · https://cyberscoop.com/npm-supply-chain-compromise-brian-fox-sonatype-op-ed [2] Bypassing AI Agent Defenses With Lies-In-The-Loop Checkmarx · Research · https://checkmarx.com/zero-post/bypassing-ai-agent-defenses-with-lies-in-the-loop [3] Early Approaches to Adversarial Fine-Tuning for Prompt Injection Defense: A 2022 Study of GPT-3 and Contemporary Models arXiv · Research · https://arxiv.org/abs/2509.14271 SAST [4] CodeCureAgent: Automatic Classification and Repair of Static Analysis Warnings arXiv · Research · https://arxiv.org/abs/2509.11787 SCA [5] AI-Powered Villager Pen Testing Tool Hits 11,000 PyPI Downloads Amid Abuse Concerns The Hacker News · Analysis · https://thehackernews.com/2025/09/ai-powered-villager-pen-testing-tool.html DAST [6] Risky Biz Soap Box: runZero shakes up vulnerability management Risky Business · Analysis · https://risky.biz/soapbox100 REMEDIATION [7] Microsoft says Windows September updates break SMBv1 shares BleepingComputer · Analysis · https://bleepingcomputer.com/news/microsoft/microsoft-says-windows-september-updates-break-smbv1-shares TRUST [8] Shaping Global AI Governance at the 80th UNGA Partnership on AI · Analysis · https://partnershiponai.org/shaping-global-ai-governance-at-the-80th-unga MODELS [9] Poison to Detect: Detection of Targeted Overfitting in Federated Learning arXiv · Research · https://arxiv.org/abs/2509.11974 [10] Reasoned Safety Alignment: Ensuring Jailbreak Defense via Answer-Then-Check arXiv · Research · https://arxiv.org/abs/2509.11629 AI SECURITY [11] The Risks of Code Assistant LLMs: Harmful Content, Misuse and Deception Palo Alto Networks · Analysis · https://unit42.paloaltonetworks.com/code-assistant-llms ADVISORIES IN THIS WINDOW 121 npm advisories · 107 critical · 10 high · 4 moderate 10 Go advisories · 3 critical · 3 high · 3 moderate · 1 low 8 Maven advisories · 6 moderate · 2 low 4 crates.io advisories · 3 high · 1 moderate 2 PyPI advisories · 1 critical · 1 moderate 1 Packagist advisory · 1 moderate 1 Hex 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.