SkillScreen: High-Recall Screening of Malicious Agent Skills with Decision Models
Agent skills turn natural-language instructions into actions, creating a security challenge: legitimate automation and malicious behaviour can look deceptively similar. A practical screen must catch nearly every attack while keeping the review load manageable. We present SkillScreen, a compact decision model trained to score skill maliciousness directly from SKILL$.$md in a single forward pass. On MaliciousSkillBench, SkillScreen flags 20-25% of benign skills at 99% recall on unseen attack families, substantially fewer than the word-level classifier and prompted LLM judges tested. Under the same training protocol, it also generalizes better than a fully fine-tuned encoder to unseen source...
| Authors | Jerrold Huang, Junwei Quan, Saiyue Lyu et al. |
|---|---|
| Primary category | cs.CR |
| Published | 13 October 2026 |
| Last updated at source | 14 October 2026 |
Recent changes
- 15 Oct 2026 Description: Agent skills turn natural-language instructions into actions, creating a security challenge: legitimate automation and malicious behaviour can look deceptively similar. A practical screen must catch nearly every attack while keeping the rev → Agent skills turn natural-language instructions into actions, creating a security challenge: legitimate automation and malicious behaviour can look deceptively similar. A practical screen must catch nearly every attack while keeping the rev
Changes detected by Silytics between refreshes of the source.
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 15 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.