Discover the toolingbehind modern intelligence
The operating system for AI infrastructure discovery. Search a live map of MCP servers, AI agents, LLM tools, automation systems, and developer infrastructure.
claude-code-book
《御舆：解码 Agent Harness》42万字拆解 AI Agent 的Harness骨架与神经 —— Claude Code 架构深度剖析，15 章从对话循环到构建你自己的 Agent Harness。在线阅读网站：
3.5kstars
738forks
agentagent-architectureai-agentanthropicbookchineseclaude-codellmmcp
Quality signals
ActiveNo testsNo CIThin docsNo licensePopular
AI summary
Auto-generated overview of what this skill does and when to reach for it.
Deep architectural analysis of AI Agent Harness systems like Claude Code, teaching how to build and understand the runtime framework powering intelligent agents.
Use this when
- Debugging why an agent stops mid-task
- Designing a new agent framework from scratch
- Evaluating tool permissions in an agent system
- Optimizing context window usage in long conversations
- Building a multi-agent coordination system
Don't use when
- Need step-by-step agent usage tutorials
- Looking for prompt engineering tricks
- Want to integrate a specific API without understanding its internals
Capabilities
- Explains 50+ design decisions behind agent runtime frameworks
- Provides 139 architecture diagrams for reusable patterns
- Covers tool systems, memory, context compression, and lifecycle hooks
- Analyzes concurrency, streaming, and error handling patterns
- Includes 89 feature flags and 50+ tool examples
- Teaches how to build your own Agent Harness from scratch
Install
Universal installANY TOOL
$ npx skillcity@latest lintsinghua/claude-code-bookAuto-detects the skill or MCP server in this repo and installs it for Claude Code, Codex, Cursor or Cline.
Requires Node.js 18+ (ships with npx).
Prefer to set it up by hand? Clone the repository:
git clone https://github.com/lintsinghua/claude-code-book.git
Loading README…
Ask a simulated agent to run this skill. It loads the real SKILL.md and reasons over your task with NVIDIA.
claude-code-book — sandboxAI sandbox
Pick a quick prompt below or type your own — watch the agent execute this skill step by step.
guest@skillcity:~$Try a prompt