SkillLens: Adaptive Multi-Granularity Skill Reuse for Cost-Efficient LLM Agents

TL;DR

SkillLens introduces a hierarchical, multi-granularity skill reuse framework for LLM agents, enabling cost-efficient and adaptive skill retrieval and refinement across tasks.

Contribution

It proposes a novel hierarchical skill graph and retrieval method that adaptively reuses and refines skills at multiple granularities, improving efficiency and performance.

Findings

Achieves up to 6.31 percentage-point improvement in bug localization accuracy.

Raises agent success rate from 45.00% to 51.31% in ALFWorld.

Provides theoretical analysis of cost and convergence properties.

Abstract

Skill libraries have become a practical way for LLM agents to reuse procedural experience across tasks. However, existing systems typically treat skills as flat, single-resolution prompt blocks. This creates a tension between relevance and cost: injecting coarse skills can introduce irrelevant or misleading context, while rewriting entire skills is expensive and often unnecessary. We propose SkillLens, a hierarchical skill-evolution framework that organizes skills into a four-layer graph of policies, strategies, procedures, and primitives, and retrieves them at mixed granularity. Given a task, SkillLens first retrieves semantically relevant skill seeds, expands them through degree-corrected random walk over the skill graph, and then uses a verifier to decide whether each visited unit should be accepted, decomposed, rewritten, or skipped. This enables the agent to reuse compatible subskills directly while adapting only locally mismatched components. To improve the system over time, SkillLens further refines multi-granularity skills and verifier in order to improve its routing decisions. We provide theoretical analysis showing that mixed-granularity adaptation incurs sublinear cost under sparse mismatch assumptions and that the evolutionary update rule monotonically improves the validation objective until a local optimum. Across MuLocbench and ALFWorld, SkillLens consistently improves over strong skill-based baselines, achieving up to a 6.31 percentage-point Acc@1 gain for bug localization and raising agent success rate from 45.00% to 51.31%.