Paper 2604.08377
SkillClaw: Let Skills Evolve Collectively with Agentic Evolver
- Published
- Apr 2026
- Research lab
- Independent
- Citations
- 48
- GitHub
- 2.4K stars
01 In brief
Summary
SkillClaw is a framework for collective skill evolution in multi-user LLM agent ecosystems, addressing the problem of static skills that do not improve with experience.
It aggregates interaction trajectories from multiple users, groups them by referenced skills, and uses an agentic evolver to refine existing skills, create new ones, or skip updates based on evidence.
Updated skills are validated nightly in real environments and synchronized back to all agents, forming a closed loop of interaction, evidence, evolution, validation, and deployment.
Experiments on WildClawBench with Qwen3-Max over 6 days show consistent improvements across four categories: Social Interaction (+11.72% relative), Search & Retrieval (+52.00%), Creative Synthesis (+88.41%), and Safety & Alignment (+33.33%).
Controlled validation on three custom queries showed an average gain of +42.1%.
The framework enables cross-user knowledge transfer and cumulative capability improvement without user effort, and is compatible with various Claw-style agent systems.
Key properties include collective evolution, full automation, and agentic adaptability, distinguishing it from memory-based or static skill-based approaches.
The paper includes case studies demonstrating improvements in task decomposition, error correction, and constraint-aware reasoning, and discusses related work in agent self-evolution and agent skills.
SkillClaw represents a shift from static skill libraries to dynamic, interaction-driven skill ecosystems, with potential for further gains with more users and longer time horizons.
The framework is open-sourced on GitHub.
The paper is a work in progress, with additional benchmark categories to be included in future versions.
The validation mechanism ensures monotonic deployment behavior, as only accepted updates are deployed, preventing degradation over…
02 From the paper
Abstract
Large language model (LLM) agents such as OpenClaw rely on reusable skills to perform complex tasks, yet these skills remain largely static after deployment. As a result, similar workflows, tool usage patterns, and failure modes are repeatedly rediscovered across users, preventing the system from improving with experience. While interactions from different users provide complementary signals about when a skill works or fails, existing systems lack a mechanism to convert such heterogeneous experiences into reliable skill updates. To address these issues, we present SkillClaw, a framework for collective skill evolution in multi-user agent ecosystems, which treats cross-user and over-time interactions as the primary signal for improving skills. SkillClaw continuously aggregates trajectories generated during use and processes them with an autonomous evolver, which identifies recurring behavioral patterns and translates them into updates to the skill set by refining existing skills or extending them with new capabilities. The resulting skills are maintained in a shared repository and synchronized across users, allowing improvements discovered in one context to propagate system-wide while requiring no additional effort from users. By integrating multi-user experience into ongoing skill updates, SkillClaw enables cross-user knowledge transfer and cumulative capability improvement, and experiments on WildClawBench show that limited interaction and feedback, it significantly improves the performance of Qwen3-Max in real-world agent scenarios.