Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems

2026-04-14

AI Summary

This work proposes an AI agent system architecture that holistically integrates human values, safety, and capability expansion. By analyzing the open-source implementation of Claude Code and comparing it with the OpenClaw system, the study systematically maps human value requirements onto thirteen design principles and identifies six critical open research directions. The architecture employs a modular TypeScript design featuring seven-mode permission control, ML-based classifiers, five-layer context compression, and four extensibility mechanisms—namely MCP, plugins, skills, and hooks—alongside sub-agent isolation with delegated authority and append-only conversation storage. Furthermore, the research delineates differentiated design pathways across multiple application scenarios, offering a reusable framework and empirical foundation for future agent systems.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent/AI Theories and Architectures
• Humans and AI: Learning Human Values and Preferences

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

Claude Code is an agentic coding tool that can run shell commands, edit files, and call external services on behalf of the user. This study describes its comprehensive architecture by analyzing the publicly available TypeScript source code and further comparing it with OpenClaw, an independent open-source AI agent system that answers many of the same design questions from a different deployment context. Our analysis identifies five human values, philosophies, and needs that motivate the architecture (human decision authority, safety and security, reliable execution, capability amplification, and contextual adaptability) and traces them through thirteen design principles to specific implementation choices. The core of the system is a simple while-loop that calls the model, runs tools, and repeats. Most of the code, however, lives in the systems around this loop: a permission system with seven modes and an ML-based classifier, a five-layer compaction pipeline for context management, four extensibility mechanisms (MCP, plugins, skills, and hooks), a subagent delegation mechanism with worktree isolation, and append-oriented session storage. A comparison with OpenClaw, a multi-channel personal assistant gateway, shows that the same recurring design questions produce different architectural answers when the deployment context changes: from per-action safety classification to perimeter-level access control, from a single CLI loop to an embedded runtime within a gateway control plane, and from context-window extensions to gateway-wide capability registration. We finally identify six open design directions for future agent systems, grounded in recent empirical, architectural, and policy literature.

Problem

• AI agent systems
• design space
• deployment context
• human values
• system architecture

Innovation

• AI agent architecture
• permission system
• context compaction pipeline
• subagent delegation
• extensibility mechanisms