Summary
WikiSkill (arXiv:2508.11371) is a framework that co-evolves agent skills with a persistent knowledge base (wiki). Agent skills package specialized knowledge and workflows into reusable resources, and recent research automatically discovers them from agent experience. However, insights guiding skill development usually remain scattered across optimization histories, limiting systematic reuse. WikiSkill addresses this by separating raw execution experience, accumulated knowledge, and executable skills, while continuously consolidating experience into a wiki on which subsequent skill updates build. Across diverse benchmarks and models, WikiSkill consistently outperforms state-of-the-art skill evolution methods and beats skill-free baselines in most settings. The authors find skill evolution is complementary to model scaling: larger models benefit more from evolved skills, and smaller models with skills can outperform larger models without them. Evolved skills also transfer effectively across models and model families. Ablation studies confirm that accumulating persistent knowledge in the wiki is essential for effective skill evolution.
Paper Overview
Field: NLP
Authors: Liyan Tang, Cyrus Rashtchian, Chun-Sung Ferng
Published: 2026-08-28
arXiv: 2508.11371
Summary
Agent skills package specialized knowledge and workflows into reusable resources that extend AI agent capabilities. Recent work automatically discovers such skills from agent experience, enabling 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.
The authors 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.
Key Findings
- Across diverse benchmarks and models, WikiSkill consistently outperforms state-of-the-art skill evolution methods and beats skill-free baselines in most model-benchmark settings.
- Skill evolution is complementary to model scaling: larger models generally benefit more from evolved skills, and smaller models equipped with skills can outperform larger models without them.
- Evolved skills transfer effectively across models and model families; skills evolved by other models can outperform self-evolved ones.
- Ablation studies confirm that accumulation of persistent knowledge in the wiki is essential for effective skill evolution.
These results demonstrate the benefits of systematically accumulating and refining agent experience for developing reusable and transferable skills.
---
*Auto-collected on 2026-08-29*
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/178634189