WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution
2026-08-27 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors present WikiSkill, a system that helps AI agents learn and improve skills by storing their experiences in a shared knowledge base called a wiki. This wiki organizes past experiences so new skills can build on what was previously learned, making skill development more efficient and reusable. They found that WikiSkill helps agents perform better across many tests and that bigger AI models gain more from using evolved skills. Additionally, skills learned by one model can be useful for other models, showing good transferability. Their studies also show that having a persistent knowledge base is key to making skill evolution work well.
AI agentskill evolutionknowledge basewikimodel scalingtransfer learningreusable skillsskill discoveryexperience consolidation
Authors
Liyan Tang, Cyrus Rashtchian, Chun-Sung Ferng, Andrew Tomkins, Da-Cheng Juan, Tu Vu
Abstract
Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers such skills from agent experience, which enables agents to progressively adapt through interaction. However, the insights that guide skill development typically remain scattered across optimization histories, limiting their systematic reuse across iterations. We introduce WikiSkill, a framework that co-evolves agent skills with a persistent knowledge base (wiki). At a high level, WikiSkill separates raw execution experience, accumulated knowledge, and executable skills, while continuously consolidating experience into the wiki, which subsequent skill updates can build on. Across diverse benchmarks and models, WikiSkill consistently outperforms state-of-the-art skill-evolution methods and improves over no-skill baselines in most model-benchmark settings. We find that skill evolution complements model scaling: larger models generally benefit more from evolved skills, while smaller models with skills can outperform substantially larger models without them. We also find that evolved skills transfer effectively across models and model families, and skills evolved by other models can outperform self-evolved skills. Finally, our ablation studies confirm that persistent knowledge accumulation in the wiki is critical for effective skill evolution. These results demonstrate the benefits of systematically accumulating and refining agent experience for developing reusable and transferable skills.