简体中文 | English
This repository collects practical workflow notes for working with AI coding agents. The focus is not prompt tricks. It is a repeatable engineering process: clarify the idea, write constraints, plan the work, implement with evidence, and keep rules reusable across projects.
- How to turn a rough project idea into a clearer implementation brief.
- How to use rule files such as
AGENTS.mdorCLAUDE.mdwithout making them vague. - How to define hard constraints that AI agents are more likely to follow.
- How to structure long-running implementation work with plans.
- How to keep reusable rules under
rules/.
README.zh-CN.mdis the full Chinese guide.rules/README.mdexplains the rule files.rules/spec_driven_rules.mdrecords spec-driven development rules.rules/my_own_rules.mdrecords personal engineering rules.rules/fractal_doc_rules.mddescribes the dynamic documentation structure.rules/plans.mddescribes execution plans for long-running work.
- Start with the Chinese guide in README.zh-CN.md.
- Copy only the rule files that match the target project.
- Keep rules strict and executable: prefer
MUST,DO,DO NOT, and concrete validation. - Update the rule files when a repeated workflow becomes stable.
The guide and rule documents are licensed under Creative Commons Attribution 4.0 International. See LICENSE.