Back to directory

dg-ai-notes

A Chinese-language study guide that walks through the internals of the Pi-Agent SDK's architecture and includes a companion agent skill and a worked example for building a data-analysis agent.

dg-ai-notes

dg-ai-notes is a Chinese-language set of study notes and tutorials about Pi-Agent, an open-source agent SDK that underlies products such as OpenClaw. A ten-chapter walkthrough explains the SDK's three-layer architecture, its agent loop, tool system, message handling, and session management, with parallel TypeScript and Python code and an experimental sandbox for stepping through the agent loop.

A second, seven-chapter track builds a real vertical agent, a data-analysis assistant, from environment setup through defining tools and packaging it as a service, with runnable companion code for each lesson. The project also ships a companion Claude Agent Skill kept in sync with the tutorial that lets an AI assistant look up Pi-Agent SDK usage and generate working configuration or code on request.

The material is available to read online, as downloadable Markdown, or as a PDF, and is licensed under the MIT license for code and CC-BY-SA-4.0 for the writing.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.