The Scaling Laws of Skills in LLM Agent Systems

TL;DR

This paper uncovers scaling laws governing skill accumulation in large language model agent systems, revealing how library size impacts routing accuracy and downstream performance.

Contribution

It introduces two coupled laws describing skill library scaling effects and demonstrates how these laws can guide optimization to improve agent robustness and efficiency.

Findings

Routing accuracy decays logarithmically with library size.

Correct execution can improve downstream decisions by about 4×.

Law-guided optimization significantly enhances routing and execution performance.

Abstract

As agent systems scale, skills accumulate into large reusable libraries, yet their scaling laws remain poorly understood. Across 15 frontier LLMs, 1,141 real-world skills, and over 3M routing or execution decisions, we identify two coupled laws. Routing law: single-step routing accuracy decays logarithmically with library size ($R^2{>}0.97$ for all models), with errors progressing from local skill competition to cross-family drift and capture by overly general "black-hole skills". Execution law: before state realization, joint routing is approximately multiplicative, whereas correct execution can improve difficult downstream decisions by about $4{\times}$. A single parameter, the routing logarithmic decay slope $b$, couples the two laws: routing-side fits predict execution-side rescue across models, showing that the same library property controls both pre-execution collapse and downstream recoverability. The laws are actionable: law-guided optimization raises held-out routing accuracy from 71.3% to 91.7%, reduces hijack from 22.4% to 4.1%, and transfers directionally to downstream ClawBench and ClawMark execution settings, improving mean pass rate from 49.3% to 61.6% on ClawBench and from 28.4% to 34.5% on ClawMark. These results show that agent performance depends not only on model capability, but also on the structure, granularity, and exposure policy of the skill library.