Where Natural Language Ends in Agents: Executable Skill Composition with MetaSkill

Meng Shu; Wei He; Hang Zhou; Runke Liu; Kai Han; Yunhe Wang

cs.AIOct 17, 2026

Where Natural Language Ends in Agents: Executable Skill Composition with MetaSkill

Authors: Meng Shu, Wei He, Hang Zhou, Runke Liu, Kai Han, Yunhe Wang

Organizations: TokenRhythm Technologies · Xidian University · Beijing Institute of Technology

Abstract

Agent Skills express open-ended task knowledge in natural language, but composing them requires persistent execution order and upstream-output handoffs. We propose selective formalization at Skill boundaries: coordination relations with stable references and machine-checkable constraints become executable, while substantive task judgments remain within Skills. We introduce MetaSkill, a hybrid representation that preserves complete Skill packages and local instructions while making invocation dependencies and predecessor-text bindings executable through a directed acyclic graph (DAG) runtime. This interface supports explicit coordination without prescribing how each Skill performs its local task. Across 20 multi-Skill artifact tasks and two LLMs, we compare natural-language procedures, prompt-visible graphs, and MetaSkill derived from a common frozen plan per task, using the same designated Skill packages. Additional comparisons evaluate AgentSkillOS with the same available Skill packages as MetaSkill, as well as agent systems using their configured tools. Blinded artifact judgments yield favorable preference point estimates across all reported comparisons, with 95% task-bootstrap intervals excluding parity against Prompt Graph under DeepSeek-V4-Flash and OpenClaw under GLM-5.2. In audited GLM-5.2 execution traces, all planned nodes complete and all declared precedence constraints are satisfied. Together, these findings support MetaSkill as a practical interface for composing open-ended Skills, combining local linguistic flexibility with executable dependencies, explicit predecessor-text bindings, and auditable coordination. Code is available at: https://github.com/TokenRhythm/opensquilla/tree/paper/metaskill-lp4fm-2026

Figures & tables

Panel of Figure 1 — Figure 1: MetaSkill selectively formalizes coordination across Skill invocations. (a) Nodes invoke complete Skill packages via $\kappa_{i}$ , with local instructions in $L_{i}$ and executable coordination $\Gamma_{i}$ . The binding supplies $v_{2}$ ’s text to $v_{4}$ ; files remain in a shared workspace. (b) After one structural validation, each runtime scheduling step computes readiness, renders bindings, dispatches at most $K$ ready nodes, and stores outputs to release dependents.

Panel of Figure 2 — Figure 2: One task, three coordination interfaces. NL-SOP and Prompt Graph coordinate within the host agent; MetaSkill separates runtime coordination from model-driven Skill execution. All share the frozen task, complete Skills, objectives, expected outputs, backbone, and host harness. Node IDs are case-local; canonical Skill names and abridged interfaces are shown. The illustration omits the shared upstream planning node and text-truncation filters. MetaSkill’s depends_on lists prerequisites, with contains invocation arguments, and outputs.v2 and outputs.v3 reference stored predecessor text.

Backbone	Baseline	W/T/L	Preference $\uparrow$	95% task CI
GLM-5.2	NL-SOP	23/22/15	0.567	[0.433, 0.692]
	Prompt Graph	22/22/16	0.550	[0.425, 0.667]
DeepSeek-V4-Flash	NL-SOP	22/23/15	0.558	[0.467, 0.650]
	Prompt Graph	28/23/9	0.658	[0.575, 0.742]

Table 1: Coordination-interface comparison: blinded pairwise artifact preference.

Setting	W/T/L	Preference $\uparrow$	95% task CI
Matched Skills	26/13/21	0.542	[0.425, 0.650]

Table 2: MetaSkill versus AgentSkillOS with matched Skills: blinded pairwise artifact preference.

Comparator	W/T/L	Preference $\uparrow$	95% task CI
OpenSquilla (NL-SOP)	23/22/15	0.567	[0.433, 0.692]
OpenClaw	27/21/12	0.625	[0.525, 0.725]
Hermes Agent	25/19/16	0.575	[0.433, 0.717]

Table 3: Agent-stack comparison under GLM-5.2: blinded pairwise artifact preference.

LLM Agent Orchestration

Large language model agent orchestration coordinates models, tools, and workflows to complete complex tasks reliably and efficiently.

See all 458 papers

SoK: Agentic Skills -- Beyond Tool Use in LLM Agents
Feb 24, 2026Yanna Jiang, Delong Li, Haiyu Deng +4Agentic AI LLM Agent Skill Learning
University of Technology Sydney · CSIRO Data61
Generative Skill Composition for LLM Agents
Jun 30, 2026Xinyu Zhao, Zhen Tan, Vaishnav Tadiparthi +5AI Coding Agents LLM Agent Skill Learning
University of North Carolina at Chapel Hill · Arizona State University · Honda Research Institute USA
SkillFusion: Towards Reliable Multi-Skill Composition for LLM Agents
Oct 11, 2026Enhao Huang, Xinye Zhao, Yiming Li +5LLM Agents LLM Agent Skill Learning
Zhejiang University · Nanjing University · Nanyang Technological University +1

Feb 24, 2026J/K move · Enter open · S save

SoK: Agentic Skills -- Beyond Tool Use in LLM Agents

Yanna Jiang, Delong Li, Haiyu Deng, Baihe Ma+3

University of Technology Sydney · CSIRO Data61

Agentic AI LLM Agent Skill Learning LLM Agent Security

Open paper AskPDF arXiv

Agentic systems increasingly rely on reusable procedural capabilities, a.k.a., agentic skills, to execute long-horizon workflows reliably. These capabilities are callable modules that package procedural knowledge with explicit applicability conditions, execution policies, termination criteria, and reusable interfaces. Unlike one-off plans or atomic tool calls, skills operate (and often do well) across tasks. This paper maps the skill layer across the full lifecycle (discovery, practice, distillation, storage, composition, evaluation, and update) and introduces two complementary taxonomies. The first is a system-level set of seven design patterns capturing how skills are packaged and executed in practice, from metadata-driven progressive disclosure and executable code skills to self-evolving libraries and marketplace distribution. The second is an orthogonal \textbf{representation

\times

scope} taxonomy describing what skills are (natural language, code, policy, hybrid) and what environments they operate over (web, OS, software engineering, robotics). We analyze the security and governance implications of skill-based agents, covering supply-chain risks, prompt injection via skill payloads, and trust-tiered execution, grounded by a case study of the ClawHavoc campaign in which nearly 1{,}200 malicious skills infiltrated a major agent marketplace, exfiltrating API keys, cryptocurrency wallets, and browser credentials at scale. We further survey deterministic evaluation approaches, anchored by recent benchmark evidence that curated skills can substantially improve agent success rates while self-generated skills may degrade them. We conclude with open challenges toward robust, verifiable, and certifiable skills for real-world autonomous agents.

Abstract

LLM Agent Orchestration

Explore similar work

SoK: Agentic Skills -- Beyond Tool Use in LLM Agents

Generative Skill Composition for LLM Agents

SkillFusion: Towards Reliable Multi-Skill Composition for LLM Agents