Group of Skills: Group-Structured Skill Retrieval for Agent Skill Libraries

2026-05-07

AI Summary

Existing agents struggle to identify execution entry points, supporting skills, observable requirements, and error-avoidance strategies from flat retrieval results when leveraging large, reusable skill libraries. This work proposes a runtime grouped and structured skill retrieval method that introduces, for the first time, role-tagged skill groups to explicitly model functional roles and dependencies among skills. Through anchor construction, group-graph expansion, and execution contract rendering, the approach generates typed skill graphs that produce fixed execution contracts comprising Start, Support, Check, and Avoid fields. Evaluated on SkillsBench and ALFWorld, the method demonstrates significantly higher coverage of observable requirements, improved task rewards, and greater runtime efficiency than flat retrieval baselines under constrained skill budgets.

Abstract

Skill-augmented agents increasingly rely on large reusable skill libraries, but retrieving relevant skills is not the same as presenting usable context. Existing methods typically return atomic skills or dependency-aware bundles whose internal roles remain implicit, leaving the agent to infer the execution entry point, support skills, visible requirements, and failure-avoidance guidance. We introduce Group of Skills (GoSkills), an inference-time group-structured retrieval method that changes the agent-facing retrieval object from a flat skill list to a compact, role-labeled execution context. GoSkills builds anchor-centered skill groups from a typed skill graph, expands support groups through a group graph, bottlenecks the selected group plan into a bounded set of atomic skill payloads, and renders a fixed execution contract with Start, Support, Check, and Avoid fields, without changing the downstream agent, skill payloads, or execution environment. Experiments on SkillsBench and ALFWorld show that GoSkills preserves visible-requirement coverage under a small skill budget, improves over flat skill-access baselines, and often improves reward and agent-only runtime relative to structural retrieval references.

Problem

• skill retrieval
• agent skill libraries
• execution context
• role-labeled skills
• skill composition

Innovation

• Group-Structured Retrieval
• Skill Libraries
• Execution Context
• Role-Labeled Skills
• Inference-Time Retrieval