Battle-tested agent rules, skills, and commands for AI coding assistants

This repository provides a comprehensive collection of production-ready configurations for AI coding assistants like OpenCode, Cursor, and Claude Code.

• core: Essential rules, skills, and commands for any project
  • TDD Best Practices
  • Git Workflow
  • Playwright Skill
  • Shell Executor
  • Test Suite Command
  • Lint All Command

• golang: Go development guidelines and conventions

  • Senior Go guidelines (error handling, concurrency, package structure)
  • Testify usage patterns
  • Interface design principles

• node: Node.js and TypeScript/JavaScript guidelines

  • TypeScript conventions
  • Type safety patterns
  • Async/await best practices

• docker: Containerization best practices
  • Production-ready Docker patterns
  • Multi-stage builds
  • Security guidelines

Each plugin follows this structure:

plugins/
└── <plugin-name>/
    ├── plugin.yaml              # Plugin manifest
    └── templates/
        ├── rules/              # Coding standards and guidelines
        ├── skills/             # AI capabilities (web, browser, etc.)
        └── commands/          # Reusable development workflows

name: <plugin-name>
version: 1.0.0
description: <plugin description>
conditions:
  - file: <detect-file>     # File that triggers plugin load
  - file: <another-file>

---
title: <rule title>
description: <what this rule enforces>
globs: ["**/*.<ext>", "**/*.<ext>"]
alwaysApply: true
source: <authoritative source URL>
references:
  - <additional reading>
  - <reference link>
---

# RULE TITLE

<rule content in markdown>

---
name: <skill-name>
description: <what this skill enables>
parameters:
  type: object
  properties:
    <param-name>: { type: string, description: "..." }
execute:
  command: "<command>"
  args: ["<arg1>", "{{.<param>}}"]
source: <reference URL>
references:
  - <additional resources>
---

# Skill Title

<skill documentation in markdown>

---
name: <command-name>
description: <what this command does>
steps:
  - name: <step name>
    command: "<command>"
---

# Command Title

<command documentation in markdown>

1. Add plugin reference to your .vendatta/config.yaml:

extends:
  - inizio/vendatta-config-inizio

2. Vendatta automatically:
   • Clones the latest plugin version
   • Applies rules based on file detection
   • Activates skills for your AI assistant
   • Makes commands available

This repository follows the Conventional Commits specification:

feat: add new plugin
fix: correct rule behavior
docs: update documentation
test: add test coverage

1. Create plugin directory: plugins/<plugin-name>/
2. Add plugin.yaml with manifest
3. Create templates in templates/rules/, templates/skills/, or templates/commands/
4. Follow the template structure above
5. Test by adding to your .vendatta/config.yaml

This repository aggregates best practices from:

• awesome-claude-code - Curated Claude Code resources
• agent-rules - Peter Steinberger's agent rules
• agent-scripts - Shared agent helper scripts
• Conventional Commits specification
• Official language style guides (Go, TypeScript, etc.)

MIT License - See LICENSE file for details