bojieli/ai-agent-book
《深入理解 AI Agent:设计原理与工程实践》(李博杰 著)开源主仓库:全书正文、编译版 PDF 与按章配套代码
01 / MACHINE INTELLIGENCE
Models you can run. Agents you can inspect. Tools you can build on.
← All topics《深入理解 AI Agent:设计原理与工程实践》(李博杰 著)开源主仓库:全书正文、编译版 PDF 与按章配套代码
he optimized, token-efficient version of the leaked Claude Fable 5 / Mythos 5 system prompt. Re-engineered into clean Markdown for universal execution on Gemini 3.1 Pro, ChatGPT 5.6, and advanced LLM agents.
Claude skill: design any process, idea, or problem into an ICM workspace (folder structure as agent architecture), or restructure an existing folder into one
Report issues related to MCP integration with Claude here.
Claude skills for founders.
Latest Advances on Agentic AI & AI Agents for Healthcare
All available LTX-2 models, encoders, workflows, LoRAs for ComfyUI
🔐 Secure, real-time monitoring dashboard for OpenClaw AI agents. Auth, TOTP MFA, cost tracking, live feed, memory browser and more.
Episodic memory for AI agents. Records your screen locally, compiles it into structured activity frames, serves them over MCP. No cloud, no LLM in the loop.
Automation harness that compiles a screen recording into a rerunnable script, with LLMs, only at decision points and self-healing reruns.
🐎 Ryan Lopopolo’s anthology, field guide, and agent context bundle for harness engineering
Container-free, deny-by-default sandbox for AI coding agents. Kernel-enforced filesystem, network, and syscall isolation for Linux and macOS
Self-evolving multi-agent "newsroom" that turns your documents into a cross-linked knowledge wiki — writer ≠ reviewer, local-first, no API keys, a structured alternative to RAG.
TreeThink: Mathematical Reasoning of LLMs with Tree Search Methods
Portable Claude Code multi-agent pipeline - skills, agents, hooks, orchestration scripts, and quality gates
Distributed agent coordination platform where agents live, connect, coordinate, execute, and evolve together.
px0 is an open-source prompt infrastructure toolkit that lets you version, update, and govern prompts in production without changing application code.
LLM agents as your hyperparameter optimizer.
Claude Code skill for free Google AI Mode search with citations. Zero-config setup, persistent browser profile, query optimization. Token-efficient web research.
[ECCV 26] FSDC-DETR.
A physics-grounded, cost-aware optimizer for vLLM.
利用claude code agent框架一步一步实现deep research!很强大很简单的skills。我一步一步介绍实现deep research,因为deep research就是agent框架第一应用,对比一下各个框架实现这个deep research,就知道哪个框架才是真厉害。
Bring any model to Codex, assign them any role, use them in /goal or any workflow.
Code for the ShapeR research paper.