AntiSkillBench¶
Product or Programme · All products and programmes
Evidence · 4 claims (4 for) · 1 study (1 design),q2· 0 of 1 report an effect size · 4 claims rest on one study
Description¶
AntiSkillBench is a benchmark dataset and evaluation suite created by Yongli Xiang et al. for assessing persona-skill privacy leakage, attribute disclosure, and behavioral impersonation risks in personalized LLM-based agent systems.
Components¶
Claims¶
- Skill-level leakage is strongest for communication style and personality, not explicit demographics [+W]
- Persona-skill distillation encodes substantial private information into skill artifacts, persisting across agent backbones and distillation protocols [+W]
- Skill-equipped agents reproduce target users' attributes and language, enabling impersonation across unseen contexts [+W]
- Active defenses (Privacy Sanitization, Adversarial Obfuscation) reduce but do not eliminate persona-skill leakage, leaving personality and background exposed [+W]
Related Products and Programmes¶
-
Key Sources¶
- Yongli Xiang, Zhifang Zhang, Bojun Yang, Ziming Hong, Lei Feng, Miao Xu, Tongliang Liu. (2026). When Agents Learn to Be You: Benchmarking Privacy Leakage, Impersonation Risk, and Defenses in Persona Skills. arXiv. https://arxiv.org/abs/2608.03700
Page metadata
| Field | Value |
|---|---|
type |
product |
id |
antiskillbench |
title |
AntiSkillBench |
description |
AntiSkillBench is a benchmark dataset and evaluation suite created by Yongli Xiang et al. |
status |
review |
product_kind |
dataset |
generated |
by process:settle-candidates on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
Frontmatter is the page's machine-readable record — see the Schema & Guide.