This repository contains reference files and rule sets for AI agents to ensure code quality, consistency, and adherence to best practices across different programming languages and frameworks.
Repository: https://github.com/gdellis/agent-files
Use the raw GitHub URLs to reference rules in your AI prompts. This ensures the agent can access the latest version of the rules:
https://github.com/gdellis/agent-files/raw/refs/heads/main/rules/<file>.md
Example Prompts:
Write a Python REST API following the rules in:
https://github.com/gdellis/agent-files/raw/refs/heads/main/rules/python.md
Review this code for bugs and style issues following:
https://github.com/gdellis/agent-files/raw/refs/heads/main/rules/rust.md
Create a bash script to deploy the application, following:
https://github.com/gdellis/agent-files/raw/refs/heads/main/rules/bash.md
Agent personas are specialized instructions for specific tasks. Reference them in your prompts:
Code Reviewer:
Review this pull request as a code reviewer following https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/code-reviewer.md
Navigation:
Find all API endpoints in this codebase following https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/navigation.md
Research:
Research best practices for WebSocket implementation following https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/research.md
Project Creator:
Create a new Rust project called 'my-cli' following https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/project-creator.md
Bootstrap (for existing projects):
Bootstrap this project with agent rules from https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/bootstrap.md
To set up AI agent rules for an existing project, ask your AI agent to:
- Detect the project type (Python, Rust, Node.js, Bash, etc.)
- Create an
AGENTS.mdfile in the project root - Include references to relevant rules using raw GitHub URLs
Example prompt:
Bootstrap this project with agent rules using https://github.com/gdellis/agent-files/raw/refs/heads/main/agents/bootstrap.md
The resulting AGENTS.md will contain:
- Project-specific commands (build, test, lint)
- Links to relevant rules (language, git, markdown)
- Code style guidelines
- Key rules summary
- Bash - Best practices for shell scripting, focusing on safety and portability.
- Git - Commit conventions, branching strategies, and git workflows.
- GitHub - PR workflows, code review, and GitHub CLI usage.
- Python - PEP 8 style, type hints, testing with pytest, and dependency management.
- Rust - Idiomatic Rust patterns, error handling, and project structure.
- Markdown - Rules for documentation consistency and
markdownlint.
- Code Reviewer - Review code changes for bugs and best practices.
- Navigation - Explore and navigate codebases efficiently.
- Research - Conduct research and gather information.
- Project Creator - Create new projects with proper structure and rules.
- Bootstrap - Set up AI agent rules for existing projects.
- Documentation - Create and maintain project documentation.
- New Project - Scaffold and initialize new projects.
.
├── AGENTS.md # Instructions for AI agents working in this repo
├── README.md # This file
├── rules/ # Code style & linting rules
│ ├── bash.md
│ ├── git.md
│ ├── github.md
│ ├── markdown.md
│ ├── python.md
│ ├── rust.md
│ └── template.md
├── agents/ # AI agent personas
│ ├── bootstrap.md
│ ├── code-reviewer.md
│ ├── navigation.md
│ ├── project-creator.md
│ └── research.md
├── guides/ # Process & workflow guides
│ ├── documentation.md
│ └── new-project.md
└── .github/
└── workflows/
└── lint.yml
Each rule file follows a standard structure defined in rules/template.md:
- Core Mandates: Non-negotiable rules (safety, strict mode).
- Style: Naming, formatting.
- Error Handling: How to manage failures.
- Testing: Testing frameworks and patterns.