agent-sh/agnix

The missing linter and lsp for AI coding assistants. Validate CLAUDE.md, AGENTS.md, SKILL.md, hooks, MCP. Plugin for all major IDEs included, with autofixes.

What it solves

AI agent configurations (like those for Claude Code, Cursor, or Copilot) often fail silently due to syntax errors or incorrect formatting. A single mistake can make an agent's skills invisible or cause it to produce incorrect results. Because different tools use different formats, a configuration that works in one tool might break in another.

How it works

Agnix uses a collection of over 440 rules—sourced from official specifications and real-world breakage patterns—to validate configuration files such as CLAUDE.md, SKILL.md, and MCP configs. It can be used via a CLI, integrated into CI/CD pipelines via GitHub Actions, or used directly in code editors like VS Code, JetBrains, Neovim, and Zed.

Who it’s for

Developers and teams building or using AI-powered coding tools and agents who want to ensure their agent instructions, skills, and tool configurations are syntactically correct and compatible across different AI platforms.

Highlights

  • Multi-tool support: Validates configurations for Claude Code, Cursor, GitHub Copilot, MCP, Gemini CLI, and more.
  • Auto-fix capabilities: Offers safe and unsafe modes to automatically correct common configuration errors.
  • Editor Integration: Provides LSP support for real-time diagnostics in popular IDEs.
  • CI/CD Ready: Includes a GitHub Action to catch broken agent configs before they are merged.