Skip to main content
Glama
TheoryofShadows

@mcpx-digital/readme-score

README.md
# @mcpx-digital/readme-score

**MCP server that scores README quality and suggests improvements.**

Honest, heuristic scoring for install/usage/license/badges/structure plus relative link checks. Local markdown files only.

## Scoring rubric (100 points)

| Category | Max | What we look for |
|----------|-----|------------------|
| Title | 5 | Top-level `#` heading |
| Description | 10 | Short prose intro |
| Install | 20 | Install heading + package-manager / clone commands |
| Usage | 20 | Usage/examples heading + fenced code |
| License | 10 | License section or clear license mention |
| Badges | 10 | Status/shield badges |
| Contributing | 5 | Contributing section or link |
| Relative links | 15 | Relative paths resolve on disk |
| Structure | 5 | Enough markdown headings |

Grades: A≥90, B≥80, C≥70, D≥60, else F.

> Heuristic for **documentation completeness** — not code quality, security, or SEO.

## Install

```bash
npx -y @mcpx-digital/readme-score
```

## Cursor `mcp.json` example

```json
{
  "mcpServers": {
    "readme-score": {
      "command": "npx",
      "args": ["-y", "@mcpx-digital/readme-score"]
    }
  }
}
```

Local clone:

```json
{
  "mcpServers": {
    "readme-score": {
      "command": "node",
      "args": ["/absolute/path/to/readme-score-mcp/index.js"]
    }
  }
}
```

## Tools

| Tool | What it does |
|------|----------------|
| `score_readme` | Full 0–100 score with category breakdown |
| `suggest_improvements` | Concrete fix list |
| `check_readme_links` | Relative link existence only |

## Example prompts

- “Score `./README.md`”
- “Suggest README improvements for this repo”
- “Which relative links in the README are broken?”

## Development

```bash
git clone https://github.com/TheoryofShadows/readme-score-mcp.git
cd readme-score-mcp
npm install
npm test
```

## License

MIT

TDQS

A3.7/5.0

Scored across 3 tools

Disambiguation4/5

The three tools are mostly easy to distinguish: score_readme produces an overall score, suggest_improvements produces recommendations, and check_readme_links verifies link resolution. However, score_readme already includes relative link validity in its scoring, so an agent could be uncertain whether to call score_readme or check_readme_links for link checking.

Naming Consistency5/5

All tool names follow a consistent lowercase snake_case verb-object pattern: score_readme, suggest_improvements, check_readme_links. The naming makes the action and target predictable across the entire set.

Tool Count5/5

Three tools is well-scoped for a focused README-scoring server. Each tool covers a distinct stage of the workflow—evaluating, suggesting improvements, and validating links—without unnecessary bloat.

Completeness5/5

The toolkit covers the complete local README workflow: scoring, receiving concrete improvement suggestions, and checking relative link integrity. Since the server explicitly targets local README files, there are no major missing operations within its stated scope.

Maintenance

ActivityMaintained
ResponsivenessNo issues