1
Deterministic Replay for AI Agent Systems
探索AI代理系统中的确定性重放技术，解决大语言模型场景下的可复现与调试难题。
arXiv:2607.16200v1 Announce Type: new Abstract: AI agent systems that couple large language models (LLMs) with external tools and APIs are inherently …
2
Show HN: DRD – Git for distributed consensus failures"
像Git管理代码一样管理分布式共识故障，DRD让你精确回放和解释系统异常分歧。
Deterministic replay debugger for distributed systems - capture, replay, and explain why consensus diverged." github.com/citadel-labs/drd Comments URL…
3
Good to Go: The LOOP Skill Engine That Hits 99% Success and Slashes Token Usage by 99% via One-Shot Recording and Deterministic Replay
解决LLM重复任务中随机性失败与高成本痛点，LOOP引擎通过一次录制和确定性重放实现99%成功率并减少99%token消耗。
arXiv:2605.14237v1 Announce Type: new Abstract: Deploying AI agents for repetitive periodic tasks exposes a critical tension: Large Language Models (L…