# DecisionLint: Jev/Luna Lint

> DecisionLint: Jev/Luna Lint is an open-source MCP server for developer tools. Offline linter for questions sent to Jev and Luna decision models, callable by agents. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** vscode, CLI, self-hosted
- **Website:** https://letools.dev/tools/decisionlint
- **Repository:** https://github.com/nolindnaidoo/decisionlint
- **MCP Registry name:** io.github.nolindnaidoo/decisionlint
- **Categories:** developer tools, AI models and LLM tooling
- **Listing updated:** 2026-11-14

## About

Decision models answer the question you wrote, which is not always the one you meant. DecisionLint reads the questions a project sends to TypeSafe's Jev or OpenAI's Decisions API (gpt-6-luna), or to any platform serving either request shape, and flags wording measured to produce bad answers: a choice with no fallback option, a score given a map instead of a list, a type that does not exist, or a question that turns on a word with no stated line.

The MCP package runs the same engine as the VS Code extension and the command line. The README says the server exposes eight tools over stdio, started with `npx -y decisionlint-mcp`, so an agent can check the questions it writes before sending them. Linting needs no API key and makes no network calls; parts of a question built at runtime are counted as not read rather than guessed at.

Platform limits are read from the model id, a URL in the file or a target setting, for example Vercel's 255 choices or LiteLLM's 128 questions. Files in JSON, JavaScript, TypeScript, Python, Rust and Go are supported, and a few optional commands in the editor extension can ask a model to review a file using your own key.

## Features

- **Offline linting:** Checks run locally with no API key and no network calls.
- **Agent-callable tools:** The MCP server offers lint_text, lint_paths and list_rules among its tools.
- **37 documented rules:** Each rule has its own page with an example and a fix.
- **Platform-specific limits:** Request limits follow the platform named by the model id, a URL or a target setting.
- **Multi-language scanning:** Finds questions in JSON, JavaScript, TypeScript, Python, Rust and Go.
- **CI output formats:** The command line prints terminal text, JSON, GitHub annotations, SARIF or JUnit.

## Good for

- Check decision-model questions an agent has just written
- Catch missing fallback options in choice questions
- Add a lint step to pull requests that touch Jev or Luna prompts
- Verify request limits for a gateway such as OpenRouter or Vercel AI Gateway

## Install

### Claude Code (Terminal)

```sh
claude mcp add decisionlint -- npx -y decisionlint-mcp
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "decisionlint": {
      "command": "npx",
      "args": [
        "-y",
        "decisionlint-mcp"
      ]
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "decisionlint": {
      "command": "npx",
      "args": [
        "-y",
        "decisionlint-mcp"
      ]
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "decisionlint": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "decisionlint-mcp"
      ]
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "decisionlint": {
      "command": "npx",
      "args": [
        "-y",
        "decisionlint-mcp"
      ]
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "decisionlint": {
      "command": "npx",
      "args": [
        "-y",
        "decisionlint-mcp"
      ]
    }
  }
}
```

## Tools

- `lint_text`: Lints a piece of text containing decision-model questions.
- `lint_paths`: Lints questions found in files at given paths.
- `list_rules`: Lists the available lint rules.

## FAQ

### Does linting need an API key?

No. Linting runs offline with no key; only the optional ask commands in the editor extension use your own key.

### How do I start the MCP server?

Run npx -y decisionlint-mcp; it communicates over stdio.

### Which tools does it expose to agents?

The README names lint_text, lint_paths and list_rules, and says there are eight tools in total.

### What happens to questions built at runtime?

They are counted as not read, reported as DEC000, and never guessed at.

### Is DecisionLint: Jev/Luna Lint free?

Yes. DecisionLint: Jev/Luna Lint is open source, so you can use and self-host it for free.

### How do I add DecisionLint: Jev/Luna Lint to Claude Code?

Run this command in your terminal: claude mcp add decisionlint -- npx -y decisionlint-mcp

### What are the best alternatives to DecisionLint: Jev/Luna Lint?

The closest alternatives to DecisionLint: Jev/Luna Lint are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Concord MCP](https://ai-powered.com/mcp/concord-mcp): Concord MCP is a freemium MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/decisionlint-jev-luna-lint
