Add dataset card

README.md

@@ -0,0 +1,39 @@
+---
+task_categories:
+- text-classification
+tags:
+- safety
+- llm-agents
+- prompt-injection
+- long-horizon
+---
+
+# ShadowMem Dataset
+
+This repository contains data associated with the paper [Safeguarding LLM Agents against Long-Horizon Threats via Shadow Memory](https://huggingface.co/papers/2605.03228).
+
+ShadowMem is a defensive framework that maintains a dedicated, safety-focused agentic memory to distill and retain safety-critical context across an agent's full execution trajectory, and uses this shadow memory to proactively assess the risk of pending actions before they are executed.
+
+Code: https://github.com/ZJUWYH/ShadowMem
+
+### Related artifacts
+
+- PI judge model: https://huggingface.co/HuntingQuasar/pi-model
+- STAC judge model: https://huggingface.co/HuntingQuasar/stac-model
+- PI dataset: https://huggingface.co/datasets/HuntingQuasar/pi-dataset
+- STAC dataset: https://huggingface.co/datasets/HuntingQuasar/stac-dataset
+
+### Citation
+
+```bibtex
+@inproceedings{wang2026shadowmem,
+  author    = {Yuhui Wang and Tanqiu Jiang and Jiacheng Liang and Charles Fleming and Ting Wang},
+  title     = {Safeguarding {LLM} Agents against Long-Horizon Threats via Shadow Memory},
+  booktitle = {Proceedings of the 2026 ACM SIGSAC Conference on Computer and Communications Security},
+  series    = {CCS '26},
+  year      = {2026},
+  publisher = {Association for Computing Machinery},
+  doi       = {10.1145/3830454.3846601},
+  url       = {https://doi.org/10.1145/3830454.3846601}
+}
+```