AgentSPEX: An Agent SPecification and EXecution Language
Summary
AgentSPEX introduces a domain-specific language for specifying modular, interpretable LLM-agent workflows with explicit control flow, state management, and a visual editor, outperforming existing Python-coupled frameworks.
View Cached Full Text
Cached at: 04/22/26, 06:17 AM
Paper page - AgentSPEX: An Agent SPecification and EXecution Language
Source: https://huggingface.co/papers/2604.13346
Abstract
AgentSPEX is a domain-specific language and framework for creating structured, modular, and interpretable large language model agent workflows with explicit control flow and state management.
Language-model agent systems commonly rely on reactive prompting, in which a single instruction guides the model through an open-ended sequence of reasoning and tool-use steps, leavingcontrol flowand intermediate state implicit and making agent behavior potentially difficult to control. Orchestration frameworks such as LangGraph, DSPy, and CrewAI impose greater structure through explicitworkflow definitions, but tightly couple workflow logic with Python, making agents difficult to maintain and modify. In this paper, we introduce AgentSPEX, anAgent SPecification and EXecution Languagefor specifying LLM-agent workflows with explicitcontrol flowandmodular structure, along with a customizable agent harness. AgentSPEX supports typed steps, branching and loops, parallel execution, reusable submodules, and explicit state management, and these workflows execute within an agent harness that providestool access, asandboxed virtual environment, and support forcheckpointing,verification, andlogging. Furthermore, we provide avisual editorwith synchronized graph and workflow views for authoring and inspection. We include ready-to-use agents for deep research and scientific research, and we evaluate AgentSPEX on 7 benchmarks. Finally, we show through a user study that AgentSPEX provides a more interpretable and accessible workflow-authoring paradigm than a popular existing agent framework.
View arXiv pageView PDFProject pageGitHub11Add to collection
Get this paper in your agent:
hf papers read 2604\.13346
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2604.13346 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2604.13346 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2604.13346 in a Space README.md to link it from this page.
Collections including this paper1
Similar Articles
Self-Programmed Execution for Language-Model Agents
This paper introduces Self-Programmed Execution (SPE), an agent architecture where the language model generates its own orchestration program rather than relying on a fixed external harness. It presents 'Spell', a Lisp-based language enabling this self-editing and re-evaluation, demonstrating that frontier models can successfully perform agentic tasks using this method.
AgentSpec: Understanding Embodied Agent Scaffolds Through Controlled Composition
Introduces AgentSpec, a modular specification framework for systematically composing and analyzing embodied LLM agent scaffolds, revealing that performance depends on scaffold compatibility and interaction effects rather than isolated module strength.
I'd like to share an updated methodology for building agents.[P]
Introduces Spice, an open-source decision layer that acts as a 'brain' above execution agents like Claude Code and Codex, enabling context-aware task delegation and structured decision-making.
AOSpec: Action and Observation Co-Speculation for Low-Latency Agent Serving
AOSpec is a lossless framework that co-speculates actions and observations across the LLM agent-environment loop to reduce latency, achieving notable end-to-end latency reductions across various serving settings.
Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents
Agent S2 is a new compositional framework for computer use agents that achieves state-of-the-art performance on multiple benchmarks by utilizing Mixture-of-Grounding and Proactive Hierarchical Planning.