From Skill Text to Skill Structure: The Scheduling-Structural-Logical Representation for Agent Skills

2026-04-27

AI Summary

Current agent skills are predominantly represented as unstructured text, hindering efficient parsing, reuse, and reasoning, thereby limiting automated skill management. This work proposes a Schedule–Structure–Logic (SSL) tripartite representation framework, inspired by linguistic knowledge representation theories, which explicitly decouples skills into scheduling signals, execution structures, and logical actions. Leveraging large language models, a skill normalizer integrates memory organization packets, script theory, and conceptual dependency theory to automatically generate SSL representations. Experimental results demonstrate that the proposed approach significantly outperforms text-based baselines, improving Mean Reciprocal Rank (MRR) from 0.573 to 0.707 in skill discovery and macro F1-score from 0.744 to 0.787 in risk assessment, thereby substantially enhancing skill retrievability, auditability, and operability.

Technology Category

• Planning, Routing, and Scheduling: Model-Based Reasoning
• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

LLM agents increasingly rely on reusable skills, capability packages that combine instructions, control flow, constraints, and tool calls. In most current agent systems, however, skills are still represented by text-heavy artifacts, including SKILL.md-style documents and structured records whose machine-usable evidence remains embedded largely in natural-language descriptions. This poses a challenge for skill-centered agent systems: managing skill collections and using skills to support agent both require reasoning over invocation interfaces, execution structure, and concrete side effects that are often entangled in a single textual surface. An explicit representation of skill knowledge may therefore help make these artifacts easier for machines to acquire and leverage. Drawing on Memory Organization Packets, Script Theory, and Conceptual Dependency from Schank and Abelson's classical work on linguistic knowledge representation, we introduce what is, to our knowledge, the first structured representation for agent skill artifacts that disentangles skill-level scheduling signals, scene-level execution structure, and logic-level action and resource-use evidence: the Scheduling-Structural-Logical (SSL) representation. We instantiate SSL with an LLM-based normalizer and evaluate it on a corpus of skills in two tasks, Skill Discovery and Risk Assessment, and superiorly outperform the text-only baselines: in Skill Discovery, SSL improves MRR from 0.573 to 0.707; in Risk Assessment, it improves macro F1 from 0.744 to 0.787. These findings reveal that explicit, source-grounded structure makes agent skills easier to search and review. They also suggest that SSL is best understood as a practical step toward more inspectable, reusable, and operationally actionable skill representations for agent systems, rather than as a finished standard or an end-to-end mechanism for managing and using skills.

Problem

• agent skills
• skill representation
• structured representation
• natural language understanding
• skill management

Innovation

• Scheduling-Structural-Logical representation
• agent skills
• structured knowledge representation
• LLM-based normalization
• skill discovery