Papers
arxiv:2605.27760

SkillGrad: Optimizing Agent Skills Like Gradient Descent

Published on May 26
· Submitted by
Yifan Lan
on May 28
Authors:
,
,
,

Abstract

SkillGrad is a gradient-descent-inspired framework that optimizes agent skills through trajectory-level loss evidence and text-based gradients, enhancing skill reliability and performance in specialized domains.

AI-generated summary

Agent skills provide a lightweight way to adapt LLM agents to specialized domains by storing reusable procedural knowledge in structured files. However, whether downloaded from third parties or self-generated, these skills are often unreliable, incomplete, or outdated. Existing skill-evolution methods often address these deficiencies through heuristic reflections without an explicit optimization formulation. In this paper, we propose SkillGrad, a gradient-descent-inspired framework for optimizing agent skills. SkillGrad treats the skill package as a structured parameter to optimize in a gradient descent fashion: task executions provide trajectory-level loss evidence, automatic diagnoses then provide text-based gradients that indicate the correction directions. To stabilize optimization across iterations, a momentum agent accumulates recurring diagnostic patterns into a persistent memory overlay. Finally, an LLM-based patcher executes the parameter update by applying layer-aware edits to the skill package. Evaluated on SpreadsheetBench Verified and WikiTableQuestions, SkillGrad consistently outperforms training-based skill evolution baselines across two backbone LLMs, improving over the strongest training-based baseline by 6.7 percentage points on average. Ablations further show that momentum and contrastive diagnosis both contribute to the final skill quality.

Community

Paper submitter

SkillGrad views agent skills as optimizable parameters and improves them through iterative trajectory analysis, textual momentum, and skill patching. It significantly outperforms related methods without changing model weights.

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2605.27760
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2605.27760 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2605.27760 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2605.27760 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.