Agent-rules provides a framework for defining rules and instructions for agentic coding tools. It enables developers to create consistent, reusable workflows for tools like Cursor, Claude CLI, and Gemini CLI. This skill is useful for teams looking to automate repetitive coding tasks and enforce best practices across their development workflows.
git clone https://github.com/lirantal/agent-rules.gitAgent-rules is a framework that scaffolds coding instructions and best practices rules into projects for AI coding assistants including Cursor, GitHub Copilot, Claude CLI, and Gemini CLI. It provides pre-built rulesets covering secure coding practices, security vulnerability detection, and testing strategies. The tool supports interactive and CLI modes to generate customized rules, MCP (Model Context Protocol) configuration, and reusable custom commands. Teams use agent-rules to enforce consistent coding standards across their development workflows and automate repetitive coding tasks with AI assistants.
[{"step":"Identify the repetitive task to automate","action":"List the specific coding task (e.g., 'pull request review', 'dependency updates', 'test generation') and the tool you'll use (e.g., Cursor, Claude CLI).","tip":"Start with high-frequency tasks that have clear success criteria to validate the rules before expanding."},{"step":"Define the rule components","action":"Break down the task into enforceable rules (e.g., linting, security checks, performance thresholds). Use the example JSON structure as a template.","tip":"Prioritize rules that prevent common mistakes or enforce team standards. Avoid over-engineering early versions."},{"step":"Configure the execution environment","action":"Set up the tool's configuration file (e.g., `.cursor/rules.json` for Cursor or `gemini.json` for Gemini CLI) with the rules. Specify triggers (e.g., pull requests) and timeouts.","tip":"Test the rules in a staging environment first. Use the tool's dry-run or preview mode if available."},{"step":"Integrate with CI/CD","action":"Add the agent-rules configuration to your repository's CI pipeline (e.g., GitHub Actions, GitLab CI). Ensure the tool runs automatically on the specified triggers.","tip":"Monitor the first few runs to catch false positives or edge cases. Adjust rules based on feedback."},{"step":"Iterate and expand","action":"Gradually add more rules or tasks to the agent-rules configuration. Document changes in the repository's README or CONTRIBUTING guide.","tip":"Use version control for the rules file (e.g., Git) to track changes and collaborate with the team."}]
Generate security-focused coding rules for Cursor and Claude Code projects
Configure MCP servers for enhanced agentic coding capabilities across GitHub Copilot and Gemini CLI
Scaffold testing strategy guidelines and best practices into AI-assisted development workflows
Create reusable custom commands and prompts for common development tasks
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/lirantal/agent-rulesCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Define a set of agent-rules for [TOOL_NAME] to automate [SPECIFIC_TASK] in [PROJECT_NAME]. The rules should include [KEY_COMPONENTS] and enforce [BEST_PRACTICES]. Format the rules as a JSON configuration file for direct use with the tool. Example: 'Define agent-rules for Cursor to automate pull request reviews in the 'ecommerce-backend' project. The rules should include linting, security checks, and performance validation, and enforce the team's coding standards.'
```json
{
"name": "ecommerce-backend-pr-review",
"description": "Automated pull request review rules for the ecommerce-backend project",
"rules": {
"linting": {
"enabled": true,
"command": "npm run lint",
"severity": "error",
"auto-fix": true,
"ignore": ["node_modules", ".git"]
},
"security": {
"enabled": true,
"sast_tool": "semgrep",
"ruleset": "owasp-top-ten",
"fail_on": ["sql-injection", "hardcoded-secrets"],
"auto-remediate": false
},
"performance": {
"enabled": true,
"checks": [
{
"type": "database-query",
"threshold": "100ms",
"action": "flag"
},
{
"type": "api-response",
"threshold": "500ms",
"action": "warn"
}
]
},
"standards": {
"enabled": true,
"requirements": [
"Use TypeScript strict mode",
"Follow repository naming conventions",
"Include unit tests for new features",
"Document public APIs with JSDoc"
],
"auto-format": true,
"prettier_config": ".prettierrc"
},
"communication": {
"enabled": true,
"templates": {
"review_comment": "PR Review by automated agent-rules:\n- [ ] Linting passed: ✅\n- [ ] Security scan passed: ✅\n- [ ] Performance checks passed: ✅\n- [ ] Standards compliance: ✅\n\nReview the changes and approve if everything looks good."
}
}
},
"execution": {
"trigger": "pull_request",
"environment": "production",
"timeout": 300,
"retries": 2
}
}
```Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan