Initial upload of Nemesis Cyber Traces sample

README.md

@@ -0,0 +1,53 @@
+---
+license: cc-by-4.0
+task_categories:
+  - tabular-classification
+  - text-classification
+language:
+  - en
+tags:
+  - synthetic
+  - cybersecurity
+  - edr
+  - soc
+  - malware
+  - intrusion-detection
+  - adversarial-ml
+pretty_name: Solstice Nemesis Cyber Adversarial Traces
+size_categories:
+  - 1M<n<10M
+configs:
+  - config_name: main
+    data_files:
+      - split: train
+        path: data/main/train.parquet
+  - config_name: traces
+    data_files:
+      - split: train
+        path: data/traces/train.parquet
+---
+
+# Solstice Nemesis Cyber Adversarial Traces
+
+**High-fidelity synthetic cybersecurity event traces for SOC and EDR model training.** This dataset contains 2.5 million synthetic security events simulating realistic adversarial behavior, Endpoint Detection and Response (EDR) telemetry, and automated block outcomes.
+
+Built by [Solstice AI Studio](https://www.solsticestudio.ai/datasets) as a free sample of a larger commercial pack. 100% synthetic — no real network logs or proprietary exploit code.
+
+## What makes this different
+Most public cybersecurity datasets are static snapshots of IP traffic. **Nemesis** focuses on the **decision logic** of both the attacker and the defender. Each trace includes granular telemetry that would be seen by a modern EDR agent, including process execution, memory modification attempts, and behavioral blocking triggers.
+
+## What's in the box
+- **Main Event Stream (`main`):** 2.5M rows of EDR-style telemetry events.
+- **Decision Traces (`traces`):** Deep-dive traces of specific adversarial sequences that were blocked by simulated EDR logic.
+
+## Use Cases
+- **EDR Rule Benchmarking:** Test the efficacy of detection rules against high-fidelity synthetic anomalies.
+- **Cybersecurity LLM Training:** Fine-tune models to interpret security logs and explain "why" a specific process sequence was flagged as malicious.
+- **SOC Analyst Training:** Populate training environments with realistic alert sequences.
+
+## Data Provenance
+Generated using Solstice’s PhantasOS / SIMA simulation engine. The simulation models attacker "personas" with varying skill levels attempting to move laterally, escalate privileges, and exfiltrate data while navigating a simulated corporate environment.
+
+## Get the Full Pack
+Scale this dataset to 100M+ events, custom network topologies, and latest CVE-matched behavioral patterns.
+[www.solsticestudio.ai/datasets](https://www.solsticestudio.ai/datasets)

data/traces/train.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b00e415c6c0ebe0d0c3c985162adb8a693be197993e8a6af3a6f3f87fc58352
+size 31381