Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Home/Authors/Taiwei Shi

Taiwei Shi

2 indexed papers

Recent (6 mo)
2
With code
0
Influential cites
0
Benchmarked
0

Publications per year

2
26

Top categories

AI×2ML×1Crypto×1

Frequent co-authors

Jieyu Zhao2×
Zhikun Xu1×
Yu Feng1×
Jacob Dineen1×
Ben Zhou1×
Xuwei Ding1×

Research Timeline

2026
The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents

The paper introduces OS-BLIND, a benchmark demonstrating that current safety evaluations fail to detect critical vulnerabilities in computer-use agents when user instructions are benign, showing high attack success rates even for safety-aligned models.

Skill Reuse as Compression in Agentic RL

The paper proposes ReuseRL, a method that improves agent generalization in Reinforcement Learning by enforcing structural compressibility of successful agent trajectories into reusable skills.

Highlighted terms show continued research focus across papers

Papers

cs.LGcs.AIRecentMay 29, 2026

Skill Reuse as Compression in Agentic RL

Zhikun Xu, Yu Feng, Jacob Dineen, Taiwei Shi +2 more

The paper proposes ReuseRL, a method that improves agent generalization in Reinforcement Learning by enforcing structural compressibility of successful agent trajectories into reusable skills.

View →
cs.CRcs.AIRecentApr 12, 2026

The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents

Xuwei Ding, Skylar Zhai, Linxin Song, Jiate Li +5 more

The paper introduces OS-BLIND, a benchmark demonstrating that current safety evaluations fail to detect critical vulnerabilities in computer-use agents when user instructions are benign, showing high…

View →