Ai-Evaluation on CuraSechttps://curasec.metacog.co.kr/tags/ai-evaluation/Recent content in Ai-Evaluation on CuraSecHugoen-usMon, 24 Aug 2026 13:10:29 +0000Diagnostic Framework for Long-Horizon Security LLM Agent Failureshttps://curasec.metacog.co.kr/insights/2026-08-24-beyond-end-to-end-success-diagnosing-failures-in-long-horizo/Mon, 24 Aug 2026 13:10:29 +0000https://curasec.metacog.co.kr/insights/2026-08-24-beyond-end-to-end-success-diagnosing-failures-in-long-horizo/<ul> <li><strong>Engineer — Learn:</strong> Academic research introducing checkpoint-based diagnostics for multi-step security AI agents; relevant if you are building or evaluating agentic security tooling, but no immediate change to running systems is required.</li> <li><strong>SOC/IR — Skip</strong></li> <li><strong>Leader — Skip</strong></li> </ul>