Structured Security Auditing and Robustness Enhancement for Untrusted Agent Skills

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lv, Lijia, Tang, Xuehai, Wen, Jie, Han, Jizhong, Hu, Songlin
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913067267260416
author Lv, Lijia
Tang, Xuehai
Wen, Jie
Han, Jizhong
Hu, Songlin
author_facet Lv, Lijia
Tang, Xuehai
Wen, Jie
Han, Jizhong
Hu, Songlin
contents Agent Skills package SKILL.md files, scripts, reference documents, and repository context into reusable capability units, turning pre-load auditing from single-prompt filtering into cross-file security review. Existing guardrails often flag risk but recover malicious intent inconsistently under semantics-preserving rewrites. This paper formulates pre-load auditing for untrusted Agent Skills as a robust three-way classification task and introduces SkillGuard-Robust, which combines role-aware evidence extraction, selective semantic verification, and consistency-preserving adjudication. We evaluate SkillGuard-Robust on SkillGuardBench and two public-ecosystem extensions through five large evaluation views ranging from 254 to 404 packages. On the 404-package held-out aggregate, SkillGuard-Robust reaches 97.30% overall exact match, 98.33% malicious-risk recall, and 98.89% attack exact consistency. On the 254-package external-ecosystem view, it reaches 99.66%, 100.00%, and 100.00%, respectively. These results support a bounded conclusion: factorized package auditing materially improves frozen and public-ecosystem robustness, while harsher external-source transfer remains an open challenge.
format Preprint
id arxiv_https___arxiv_org_abs_2604_25109
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Structured Security Auditing and Robustness Enhancement for Untrusted Agent Skills
Lv, Lijia
Tang, Xuehai
Wen, Jie
Han, Jizhong
Hu, Songlin
Cryptography and Security
Artificial Intelligence
Agent Skills package SKILL.md files, scripts, reference documents, and repository context into reusable capability units, turning pre-load auditing from single-prompt filtering into cross-file security review. Existing guardrails often flag risk but recover malicious intent inconsistently under semantics-preserving rewrites. This paper formulates pre-load auditing for untrusted Agent Skills as a robust three-way classification task and introduces SkillGuard-Robust, which combines role-aware evidence extraction, selective semantic verification, and consistency-preserving adjudication. We evaluate SkillGuard-Robust on SkillGuardBench and two public-ecosystem extensions through five large evaluation views ranging from 254 to 404 packages. On the 404-package held-out aggregate, SkillGuard-Robust reaches 97.30% overall exact match, 98.33% malicious-risk recall, and 98.89% attack exact consistency. On the 254-package external-ecosystem view, it reaches 99.66%, 100.00%, and 100.00%, respectively. These results support a bounded conclusion: factorized package auditing materially improves frozen and public-ecosystem robustness, while harsher external-source transfer remains an open challenge.
title Structured Security Auditing and Robustness Enhancement for Untrusted Agent Skills
topic Cryptography and Security
Artificial Intelligence
url https://arxiv.org/abs/2604.25109