Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

SkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill Graphs

About

Skill libraries enable large language model agents to reuse experience from past interactions, but most existing libraries store skills as isolated entries and retrieve them only by semantic similarity. This leads to two key challenges for compositional tasks. Firstly, an agent must identify not only relevant skills but also how they depend on and build upon each other. Secondly, it also makes library maintenance difficult, since the system lacks structural cues for deciding when skills should be merged, split, or removed. We propose SKILLGRAPH, a framework that represents reusable skills as nodes in a directed graph, with typed edges encoding prerequisite, enhancement, and co-occurrence relations. Given a new task, SKILLGRAPH retrieves not just individual skills, but an ordered skill subgraph that can guide multi-step decision making. The graph is continuously updated from agent trajectories and reinforcement learning feedback, allowing both the skill library and the agent policy to improve together. Experiments on ALFWorld, WebShop, and seven search-augmented QA tasks show that SKILLGRAPH achieves state-of-the-art performance against memory-augmented RL methods, with especially large gains on complex tasks that require composing multiple skills.

Xiaoyuan Li, Moxin Li, Keqin Bao, Yubo Ma, Wenjie Wang, Dayiheng Liu, Fuli Feng• 2026

Related benchmarks

TaskDatasetResultRank
Interactive Decision-makingAlfWorld
Overall Success Rate90.6
295
Single-hop Question AnsweringPopQA
EM48.5
186
Single-hop Question AnsweringTriviaQA
EM63.8
133
Online ShoppingWebshop
Score91.5
61
Single-hop Question AnsweringNQ
Exact Match (EM)48
60
Multi-hop Question AnsweringMuSiQue
Exact Match (EM)19.5
51
Multi-hop Question Answering2Wiki
Exact Match (EM)43.4
18
Multi-hop Question AnsweringHotpotQA
Exact Match (EM)44.7
9
Multi-hop Question AnsweringBamboogle
Exact Match (EM)72.6
9
Showing 9 of 9 rows

Other info

Follow for update