Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Libraries

2026-07-11

AI Summary

This study systematically investigates the management of dynamically evolving skill repositories in large language model agents. Based on a comprehensive review of 124 publications from 2023 to 2026, it introduces the first integrated framework that treats skill repositories as evolvable artifacts, comprising a six-dimensional skill taxonomy, an eight-stage lifecycle architecture, and a ten-operator provenance vocabulary. The work uncovers the critical roles of skill admission and repair mechanisms, demonstrates how validator quality influences reinforcement learning efficacy, and identifies performance bottlenecks of flat retrieval strategies under scaling conditions. Building on these insights, the paper proposes standardized evaluation criteria for dynamic skill repositories and outlines key open challenges in the field.

Abstract

Large language model agents increasingly store reusable procedures outside the model. These reusable procedures are often called \emph{skills}: they may be code functions, natural-language instructions, SKILL.md packages, workflow graphs, or learned adapters that a future agent can retrieve and invoke. This taxonomy-driven survey asks how such skill libraries change over time. Across a $124$-paper $2023$--$2026$ audit set, we synthesize dynamic skill systems as \emph{lifecycle-managed, verified, evolving artifact stores}: agents collect evidence from interaction, propose skill updates, verify and admit candidates, organize them for retrieval and composition, repair or prune stale entries, and govern sharing through provenance and rollback. We organize the literature around three survey tools. First, a $\text{six}$-sense taxonomy distinguishes the structurally different artifacts called ``skills'' in current papers. Second, an $\text{eight}$-stage lifecycle architecture identifies the recurring design decisions behind evidence acquisition, proposal, verification/admission, storage, retrieval/composition, maintenance, distillation/portability, and governance. Third, a lightweight skill-record schema and $\text{ten}$-operator vocabulary provide common terms for comparing library updates without elevating them into a separate method contribution. Using this structure, we synthesize evidence-graded patterns with explicit caveats: admission and repair are repeatedly important, verifier quality materially affects skill-aware RL, flat retrieval can degrade as libraries grow, and current benchmarks still under-report library trajectories, usage--utility gaps, and safety surfaces. We close with concrete reporting standards and open problems for evaluating dynamic skills as changing libraries rather than static prompt or tool collections.

Problem

• dynamic skills
• skill lifecycle
• evolving skill libraries
• agent skill management
• skill taxonomy

Innovation

• dynamic skill libraries
• skill lifecycle
• taxonomy of skills
• agent skill evolution
• verified artifact stores