Link dataset card to paper

This PR updates the dataset card to include a link to the associated research paper [ClawHub Security Signals: When VirusTotal, Static Analysis, and SkillSpector Disagree](https://huggingface.co/papers/2606.01494). Linking the paper directly ensures that the "Papers" tab on the repository is populated and provides users with easy access to the formal research context.

README.md

@@ -1,48 +1,50 @@
 ---
 language:
-  - en
+- en
 license: mit
 size_categories:
-  - 10K<n<100K
+- 10K<n<100K
 task_categories:
-  - text-classification
+- text-classification
 task_ids:
-  - multi-class-classification
+- multi-class-classification
 pretty_name: ClawHub Security Signals
 tags:
-  - security
-  - llm-security
-  - agentic-ai
-  - agent-skills
-  - openclaw
-  - clawhub
-  - malware-detection
-  - static-analysis
-  - software-supply-chain
-  - skillspector
-  - owasp
-  - scanner-disagreement
-  - trust-and-safety
-  - mlcroissant
+- security
+- llm-security
+- agentic-ai
+- agent-skills
+- openclaw
+- clawhub
+- malware-detection
+- static-analysis
+- software-supply-chain
+- skillspector
+- owasp
+- scanner-disagreement
+- trust-and-safety
+- mlcroissant
 configs:
-  - config_name: default
-    data_files:
-      - split: train
-        path: data/train.jsonl
-      - split: validation
-        path: data/validation.jsonl
-      - split: test
-        path: data/test.jsonl
-      - split: eval_holdout
-        path: data/eval_holdout.jsonl
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train.jsonl
+  - split: validation
+    path: data/validation.jsonl
+  - split: test
+    path: data/test.jsonl
+  - split: eval_holdout
+    path: data/eval_holdout.jsonl
 ---
 
 # ClawHub Security Signals
 
-🦀 [**ClawHub**](https://clawhub.ai) | 📝 [**OpenClaw Blog**](https://openclaw.ai/blog/openclaw-nvidia-skill-security) | 🤗 [**Hugging Face Blog**](https://huggingface.co/blog/OpenClaw/clawhub-security-signals) | 📄 [**Pre-Print Paper**](https://openclaw.ai/publications/clawhub-security-signals.pdf)
+🦀 [**ClawHub**](https://clawhub.ai) | 📝 [**OpenClaw Blog**](https://openclaw.ai/blog/openclaw-nvidia-skill-security) | 🤗 [**Hugging Face Blog**](https://huggingface.co/blog/OpenClaw/clawhub-security-signals) | 📄 [**Paper**](https://huggingface.co/papers/2606.01494) | 📄 [**Pre-Print**](https://openclaw.ai/publications/clawhub-security-signals.pdf)
 
 **ClawHub Security Signals** is a sanitized, MIT-licensed security-signals dataset for public OpenClaw agent skills. It captures how an agent-skill registry evaluates trust, provenance, bundled code, and scanner evidence at scale.
 
+This dataset was presented in the paper [ClawHub Security Signals: When VirusTotal, Static Analysis, and SkillSpector Disagree](https://huggingface.co/papers/2606.01494).
+
 This Hugging Face dataset repository hosts 67,453 latest public ClawHub skill versions with redacted `SKILL.md` content, sanitized bundled files where present, ClawScan registry verdicts, and supporting scanner evidence from VirusTotal, static heuristic analysis, and NVIDIA SkillSpector.
 
 **Important framing:** scanner positives are evidence, not ground truth. SkillSpector findings are semantic agentic-risk advisories, not accusations or install-blocking verdicts by themselves. A ClawScan `suspicious` verdict means "review before trusting," not "malicious."
@@ -236,4 +238,4 @@ Hugging Face can expose this dataset through the Croissant metadata endpoint whe
 
 ## Licensing
 
-This dataset is released under the MIT license. ClawHub and public OpenClaw projects are released under the permissive MIT license at the time of publishing, which covers the sanitized signals and analyzed public skill content redistributed here.
+This dataset is released under the MIT license. ClawHub and public OpenClaw projects are released under the permissive MIT license at the time of publishing, which covers the sanitized signals and analyzed public skill content redistributed here.