SkillMarker protects agent skills by embedding hidden ownership marks
SkillMarker: A KGW-Based Watermarking Method for Agent Skill Ownership Verification
Cryptography and SecurityArtificial Intelligence
Summary
Agent skills are instructions that help software agents perform tasks, but sharing these skills openly risks copying or theft. The authors present SkillMarker, a method that hides a special watermark in these skills to prove ownership without changing their function. It does this by carefully rewriting the instructions to keep their meaning while adding a detectable mark. They also provide tools to embed, check, and find these watermarks to help track and protect skills when shared.
What this means in practice
- •For software developers: Embed ownership marks into reusable software agent skills to prove authorship and prevent unauthorized copying.
- •For security analysts: Detect leaked or copied agent skills within distributed systems to trace ownership and identify unauthorized use.
Authors
Alan Tang, Lifeng Wang, Gang Wu, Tao Zhang, Yantao Jia, Xiaofeng Zhu
Abstract
Agent skills encode operational procedures and domain expertise as reusable knowledge assets, yet their distribution in plaintext exposes them to unauthorized copying and leakage. However, practical mechanisms for skill ownership verification and leakage tracing remain underdeveloped. Watermark embedding before distribution offers a potential solution, but structured, compact skill instructions impose stringent requirements for factual and procedural accuracy. To address these challenges, we introduce SkillMarker, a KGW-based watermarking method that jointly targets functionality preservation and strong watermark embedding. SkillMarker unifies quality-constrained rewriting with signal-oriented candidate search. The former separates essential knowledge from flexible expression to preserve instruction fidelity; the latter explores alternative expressions to obtain a stronger watermark signal under preservation constraints. We release an open-source toolkit that integrates watermark embedding, quality auditing, and detection to support the protection of agent skills. SkillMarker targets ownership verification and leakage tracing in distributed skill ecosystems, where content may be copied in whole or in part, or incorporated into other skills.