agent-rules-books
AGENTS.md rules / skills for AI coding agents: Codex, Cursor & Claude Code. Inspired by Clean Code, Refactoring, DDD, Clean Architecture and DDIA programming books.
Agent Rules Books provides ready-to-use coding skills and AGENTS.md rule sets for AI coding assistants like Codex, Cursor, and Claude Code, distilled from 13 classic software engineering books including Clean Code, Refactoring, Domain-Driven Design, and Clean Architecture. Each rule set comes in three sizes—mini (recommended for most tasks), nano (compact for tight context budgets), and full (complete reference)—enabling developers to guide AI agents with proven design principles and best practices. The MIT-licensed Markdown files can be integrated as project rules, skills, or custom instructions to improve code quality, architecture decisions, and refactoring outcomes. Validated through refactoring experiments, the structured rules prove more effective than simply naming a book when instructing agents to apply specific engineering principles.
Key Features
Use Cases
- 01Guiding AI agents during code refactoring sessions to apply Clean Code or Refactoring principles systematically
- 02Improving AI-generated architecture by loading Clean Architecture or Domain-Driven Design rules as project memory
- 03Teaching AI coding assistants to handle legacy code safely using Working Effectively with Legacy Code techniques
- 04Enforcing data consistency and reliability patterns from Designing Data-Intensive Applications in microservice codebases
- 05Setting project-wide coding standards by combining Pragmatic Programmer and Code Complete rule sets in AGENTS.md
- 06Creating task-specific skills for domain modeling, API design review, or production resilience checks
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
agent-rules-books — FAQ
What is Agent Rules Books?+
Agent Rules Books is a collection of AI coding agent skills and AGENTS.md rule sets distilled from 13 classic software engineering books. The Markdown rule files guide AI assistants like Codex, Cursor, and Claude Code to apply proven design principles during code generation, refactoring, and architecture decisions.
How do I install Agent Rules Books in my coding agent?+
Clone the repository, choose a rule set and size variant (mini, nano, or full), then copy the Markdown file into your project as AGENTS.md, CLAUDE.md, or integrate it as a Cursor rule or GitHub Copilot custom instruction. For Claude Code, you can also turn a mini rule set into a reusable skill.
Which AI coding clients work with these rules?+
The rule sets are compatible with Codex, Cursor, Claude Code, and GitHub Copilot. They are plain Markdown files that any AI coding assistant capable of reading project instructions or custom rules can use.
Do I need API keys or subscriptions to use Agent Rules Books?+
No API keys are required for the rule sets themselves. You only need access to an AI coding assistant that supports project-level instructions, which may require its own subscription (e.g., Claude Pro, GitHub Copilot, Cursor Pro).
Is Agent Rules Books free to use?+
Yes, Agent Rules Books is released under the MIT License and is completely free to use, modify, and distribute.
Which rule set size should I use?+
Use mini for most real tasks (28-47 rules, ~4-8 KB), nano for very tight context budgets (14-26 rules, ~1-3 KB), and full as a complete reference (158-523 rules, ~11-62 KB). Mini is recommended as the default working version.
How do I install agent-rules-books?+
Open the source repository on GitHub and follow its README. agent-rules-books is a skill — MCP Agents Market links you directly to the official repo.
Is agent-rules-books free?+
agent-rules-books is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.