solsticestudioai/nemesis-cyber-adversarial-traces

Solstice Nemesis Cyber Adversarial Traces是一个高保真度的合成网络安全事件跟踪数据集，用于SOC（安全运营中心）和EDR（端点检测与响应）模型训练。这个公开样本包含500条模拟对抗行为、EDR遥测和自动阻止结果的合成安全跟踪记录。数据集的特点是关注攻击者和防御者的决策逻辑，每条跟踪记录包括现代EDR代理会看到的详细遥测数据，如进程执行、内存修改尝试和行为阻止触发。使用案例包括EDR规则基准测试、网络安全LLM训练和SOC分析师培训。数据是通过Solstice的PhantasOS/SIMA模拟引擎生成的，模拟了不同技能水平的攻击者试图横向移动、提升权限和窃取数据的行为。

Solstice Nemesis Cyber Adversarial Traces is a high-fidelity synthetic cybersecurity event traces dataset for SOC and EDR model training. This public sample contains 500 synthetic security trace rows simulating adversarial behavior, Endpoint Detection and Response (EDR) telemetry, and automated block outcomes. The dataset focuses on the decision logic of both the attacker and the defender, with each trace including granular telemetry that would be seen by a modern EDR agent, such as process execution, memory modification attempts, and behavioral blocking triggers. Use cases include EDR rule benchmarking, cybersecurity LLM training, and SOC analyst training. The data is generated using Solstice’s PhantasOS / SIMA simulation engine, modeling attacker personas with varying skill levels attempting to move laterally, escalate privileges, and exfiltrate data.