From Prompt Engineering to Agent Development: A Comprehensive Open-Source Guide to AI Prompt Engineering

This bilingual open-source guide to AI prompt engineering, created by Hsiu-Chi Tsai, covers everything from basic frameworks to cutting-edge 2025-2026 applications (such as Vibe Coding and AI Agents). Built with Docusaurus, it helps users master skills for collaborating with large language models and reflects the industry's shift from prompt engineering to context engineering and agent-driven workflows.

• Original author/maintainer: Hsiu-Chi Tsai (@thc1006)
• Source platform: GitHub
• Release date: June 2, 2026
• Open-source license: Apache License 2.0
• Project positioning: A bilingual (Traditional Chinese/English) open-source guide aiming to cover a complete learning path from basic concepts to advanced applications.

Basic Section

Covers core concepts of prompt engineering, basic structures, common patterns (zero-shot, few-shot, chain of thought, etc.), and a quick start guide.

Advanced Section

Highlights include Vibe Coding and AI Agents: provides hands-on manuals for Claude Code/OpenAI Codex, including writing project context files, designing agent workflows, specification-driven development, etc.

Practical Tutorials

Covers scenarios like content creation, code generation, data analysis, etc.

Advanced Techniques

Includes skills like chain of thought, prompt chaining, multimodality, etc.

Best Practices

Covers test optimization, production deployment, safety ethics, team collaboration norms.

Application Cases

Shows business intelligence analysis and enterprise-level deployment practices.

Built with Docusaurus 3, its advantages include:

• Built-in internationalization support: Switch between Chinese and English versions (English docs are in docs/, Traditional Chinese in i18n/zh-TW/)
• Responsive design: Adapts to desktop and mobile devices
• Built-in support for search, version control, and local hot-reload development.

The guide reflects industry shifts:

1. Prompt Engineering: Focuses on the quality of individual prompts
2. Context Engineering: Emphasizes providing complete project backgrounds (e.g., CLAUDE.md/AGENTS.md) to maintain task coherence
3. Agent Workflow: AI autonomously plans multi-step tasks and calls tools to collaboratively achieve complex goals (e.g., Claude Code/OpenAI Codex).

• Beginners: Start with the Basic Section + accumulate experience through Practical Tutorials
• Developers: Focus on the Vibe Coding and AI Agent chapters
• Product managers/business professionals: Refer to Application Cases and Best Practices
• Researchers: Dive into the Advanced Techniques section.

Participate via GitHub Issues/PR:

• Report bugs, suggest new topics, submit translation improvements, share cases
• Need to maintain consistency between Chinese and English versions (modifications to docs/ require updating corresponding content in i18n/zh-TW/).

The guide is both comprehensive and up-to-date, embodying the spirit of open-source knowledge sharing, lowering the threshold for AI technology, and suitable for all groups from beginners to senior practitioners to gain insights and guidance.