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 1 March 2025 Retrospective Covers 28 February 2025 00:00 to 1 March 2025 00:00 UTC Web: https://keryn.dev/headlines/daily/2025-03-01/ TOP DEVELOPMENTS [1] Microsoft Exposes LLMjacking Cybercriminals Behind Azure AI Abuse Scheme The Hacker News · Incident · https://thehackernews.com/2025/02/microsoft-exposes-llmjacking.html [2] UDora: A Unified Red Teaming Framework against LLM Agents by Dynamically Hijacking Their Own Reasoning arXiv · Research · https://arxiv.org/abs/2503.01908 [3] Steering Dialogue Dynamics for Robustness against Multi-turn Jailbreaking Attacks arXiv · Research · https://arxiv.org/abs/2503.00187 SAST [4] Artemis: Toward Accurate Detection of Server-Side Request Forgeries through LLM-Assisted Inter-Procedural Path-Sensitive Taint Analysis arXiv · Research · https://arxiv.org/abs/2502.21026 DAST [5] Infosec products of the month: February 2025 Help Net Security · Analysis · https://helpnetsecurity.com/2025/02/28/infosec-products-of-the-month-february-2025 AGENTIC SDLC [6] CyLens: Towards Reinventing Cyber Threat Intelligence in the Paradigm of Agentic Large Language Models arXiv · Research · https://arxiv.org/abs/2502.20791 REMEDIATION [7] MITRE Caldera RCE vulnerability with public PoC fixed, patch ASAP! (CVE-2025–27364) Help Net Security · Analysis · https://helpnetsecurity.com/2025/02/28/mitre-caldera-rce-vulnerability-with-public-poc-cve-2025-27364 TRUST [8] Understanding the AI Act and its compliance challenges Help Net Security · Standard / policy · https://helpnetsecurity.com/2025/02/28/david-dumont-hunton-andrews-kurth-eu-ai-act-compliance MODELS [9] Approaching the Harm of Gradient Attacks While Only Flipping Labels arXiv · Research · https://arxiv.org/abs/2503.00140 [10] FC-Attack: Jailbreaking Multimodal Large Language Models via Auto-Generated Flowcharts arXiv · Research · https://arxiv.org/abs/2502.21059 [11] Efficient Jailbreaking of Large Models by Freeze Training: Lower Layers Exhibit Greater Sensitivity to Harmful Content arXiv · Research · https://arxiv.org/abs/2502.20952 ADVISORIES IN THIS WINDOW 95 npm advisories · 95 critical 1 Maven advisory · 1 high 1 crates.io advisory · 1 moderate 1 Go advisory · 1 critical 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.