SkillSmith: Co-Evolving Skills and Tools for Self-Improving Agent Systems
TL;DR
This work forms external-policy evolution under these two budgets and introduces SkillSmith, a conflict-driven learning system that improves over EvoSkill over OfficeQA and SealQA, and sustains improvement beyond SkillClaw on WildClawBench.
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.
