Teaching AI agents to share the keyboard

What it does This repo collects five progressive tutorials that turn solo AI chat sessions into structured coding workflows. It covers macOS and Windows toolchains for Claude, Gemini, and Codex CLI, then advances to giving agents external tools via MCP and Skills, orchestrating multiple agents on long tasks, and adding evaluation and security guardrails. Think of it as a curriculum for when copy-pasting from a chat window stops scaling.

The interesting bit The series deliberately wades into the unglamorous logistics of agent governance—how to divide labor so coding agents do not overwrite each other, and how to inspect or constrain their output. The author also maintains a living list of curated Skills configurations, inviting community pull requests for new setups and platform-specific bug fixes.

Key highlights

• Five chapters ranging from VS Code setup to multi-agent pipelines and safety review
• Practical coverage of MCP and Skills to extend what coding agents can actually do
• Specific strategies for multi-agent division of labor and long-task management
• A curated, community-expandable list of ten recommended Skills
• Cross-platform focus on both macOS and Windows toolchains

Caveats

• All content is in Chinese; English readers will need translation help
• This is a Markdown tutorial collection, not a runnable framework or CLI
• Includes a sponsor banner for an API proxy service at the top of the README

Verdict Worth bookmarking if you are a Chinese-reading developer trying to move from ad-hoc prompting to reproducible agent teams. Skip it if you are hunting for drop-in code libraries or English-only documentation.