SkillsMetric: Mapping the Detection Boundary of Static Analysis for Malicious Agent Skills

arXiv:2608.08468 · cs.CR, cs.AI · Submitted 2026-08-09 · Read on arXiv

Xinze Chen, Chi Zhang, Ping Ji, Yimin Liu

cs.CR, cs.AI

Submitted: 2026-08-09

Updated: 2026-08-11

Comments: 6 pages, 3 figures, 3 tables

Code: https://github.com/anthropics/skills

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Steganography — 7× in this paper · explained in 2 episodes
• AUC — 6× in this paper · explained in 2 episodes
• Composition — 4× in this paper · explained in 2 episodes
• Obfuscation — 4× in this paper · explained in 2 episodes
• Benchmark — 3× in this paper · explained in 2 episodes
• Interpretability — 2× in this paper · explained in 15 episodes
• Overfitting — 2× in this paper · explained in 2 episodes
• Procedural Knowledge — 2× in this paper · explained in 3 episodes
• Calibration — 1× in this paper · explained in 9 episodes
• Indirect Prompt Injection — 1× in this paper · explained in 2 episodes
• Prompt Injection Attacks — 1× in this paper · explained in 2 episodes

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Estimating the Social Cost of Corporate Data Breaches
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills