SkillSmith: Co-Evolving Skills and Tools for Self-Improving Agent Systems
cs.AI
Submitted: 2026-05-31
Updated: 2026-09-26
Code: https://github.com/InternLM/WildClawBench
License: http://creativecommons.org/licenses/by/4.0/
The gist: LLM agents increasingly store capabilities in external skill and tool libraries.
Terminology
Abstract
LLM agents increasingly store capabilities in external skill and tool libraries. Libraries can grow, but execution must fit a finite context window and improvement consumes a finite iteration budget. We formulate external-policy evolution under these two budgets and introduce SkillSmith. A skill ecosystem allocates context through replicator dynamics, using randomized masking and Bayesian estimation to learn skill effects and interactions. Four typed operations repair tools and move reusable logic across the skill-tool boundary, freeing context for the ecosystem. Conflict-driven learning distills failed proposals into version-scoped nogoods that guide repairs and veto repeated failing proposals. We prove utility ascent for fixed-model continuous dynamics, conditions under which lower context costs expand attainable utility, and the need for synchronized interface migration under persistent exact-match dependencies. Across five Qwen3.5 scales, SkillSmith improves over EvoSkill by up to 10.2 and 9.8 percentage points on OfficeQA and SealQA, respectively, and sustains improvement beyond SkillClaw on WildClawBench. Skills compete for context; tools are how the system makes room.
Sources
- GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
- EvoSkill: Automated Skill Discovery for Multi-Agent Systems
- Understanding the planning of LLM agents: A survey
- XSkill: Continual Learning from Experience and Skills in Multimodal Agents
- MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning
- SkillForge: Forging Domain-Specific, Self-Evolving Agent Skills in Cloud Technical Support
- SkillClaw: Let Skills Evolve Collectively with Agentic Evolver
- WebGPT: Browser-assisted question-answering with human feedback
- Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills
- SealQA: Raising the Bar for Reasoning in Search-Augmented Language Models
- Dynamic Dual-Granularity Skill Bank for Agentic RL
- SkillX: Automatically Constructing Skill Knowledge Bases for Agents
- SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning
- ReAct: Synergizing Reasoning and Acting in Language Models
- CoEvoSkills: Self-Evolving Agent Skills via Co-Evolutionary Verification
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems