# What Should a Skill Remember? Quality--Cost Trade-offs in Cost-Aware Skill Rewriting for Language Model Agents

Source: [What Should a Skill Remember? Quality--Cost Trade-offs in Cost-Aware Skill Rewriting for Language Model Agents](<https://www.sota2.com/research/paper/6a279e8f90966ade919bfc68>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-07-21T05:00:05.525Z

Authors: Qinghua Xing, Yinda Chen, Yaping Jin, Zhenhe Wu, Bohan Lin, Hang Zhou, Xinghao Chen, Hanting Chen, Zhiwei Xiong

Published: 2026-06-08T12:36:51.000Z

## Abstract

Large language model agents increasingly rely on skills: reusable procedural documents encoding workflows, tool use, implementation patterns, validation checks, and domain rules. Skill rewriting is often treated as prompt compression, but shorter skills can make agents more expensive by removing sparse operational anchors that prevent exploration, debugging, and recovery. We study skill rewriting through this economic lens. Our controlled framework profiles skill structure, rewrites skills using information-preservation strategies, and evaluates the rewrites under fixed task instructions, environments, and verifiers. Experiments on SkillsBench reveal distinct quality--cost trade-offs across strategies: API/code anchoring, workflow guarding, and rule/formula anchoring benefit different task families, with no universally dominant template. In the main held-out evaluation, the learned policy reduces total cost by 7.0% and downstream agent-token cost by 6.0%; in frozen cross-model transfer, the corresponding reductions average 14.7% and 13.7%, while verifier quality is preserved. These results position skill design as cost-aware operational knowledge engineering rather than prompt compression. Resources: https://github.com/1Reminding/Skill\_EE.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Skill-Augmented Agent Performance Evaluation on SkillsBench 86 runnable tasks (frozen-policy transfer)](<https://www.sota2.com/research/sota/skill-augmented-agent-performance-evaluation-on-skillsbench-86-runnable-tasks-frozen-policy-transfer.md>): Skill-Augmented Agent Performance Evaluation; SkillsBench 86 runnable tasks (frozen-policy transfer); Valid Task Count: 79.

- [Skill-Augmented Agent Performance Evaluation on SkillsBench 20-task balanced panel (held-out evaluation)](<https://www.sota2.com/research/sota/skill-augmented-agent-performance-evaluation-on-skillsbench-20-task-balanced-panel-held-out-evaluation.md>): Skill-Augmented Agent Performance Evaluation; SkillsBench 20-task balanced panel (held-out evaluation); Valid Task Count: 19.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2606.09421>)

[PDF](<https://arxiv.org/pdf/2606.09421.pdf>)
