# agent-seo

agent-seo grades MCP servers and other AI agents on five trust criteria, 0 to 100. Useful before you wire someone else's agent into your stack or ship your own.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · 895 installs
- Author: web3signals
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://smithery.ai/server/web3signals/agent-seo
- Updated: 2026-04-15
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/agent-seo

## What it does

Three tools, one job each. `score_agent` takes a URL pointing at an MCP server or an AI agent and rates it on five dimensions: schema quality, maintenance health, functional reliability, developer experience and ecosystem signal. You get a number between 0 and 100.

\`compare_agents` runs two candidates and reports which one is stronger category by category, so a choice between two ready-made agents stops being a coin flip. `get_fix_recommendations` turns to your own agent and hands back ranked fixes with the points each is expected to add, plus templates you can copy and references to the spec.

The endpoint is hosted at `https://agent-seo--web3signals.run.tools` and speaks HTTP. Nothing runs locally, and the address of whatever you check leaves your network for that host.

## Requirements
- Remote server: `https://agent-seo--web3signals.run.tools`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http agent-seo https://agent-seo--web3signals.run.tools
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add agent-seo --url https://agent-seo--web3signals.run.tools
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "agent-seo": {
            "url": "https://agent-seo--web3signals.run.tools"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "agent-seo": {
            "type": "http",
            "url": "https://agent-seo--web3signals.run.tools"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http agent-seo https://agent-seo--web3signals.run.tools
```

## When to use it
- vet a third-party MCP server before wiring it in
- pick between two agents built for the same job
- find what drags your own agent's score down
- get a ranked to-do list of fixes with expected point gains
- check whether your agent is discoverable at all

## Example prompts
- Score `<url>` with `score_agent` and break the result down by dimension.
- Run `compare_agents` on these two agents and tell me which one to trust.
- Get fix recommendations for our own agent — what to fix first and what it buys us.

## Similar
- [ROAS Expert free AI SEO audit](https://affpapa.org/en/nevibecode/mcp/roas-expert-free-ai-seo-audit.md) — MCP server · ★ 1 repo stars
- [harness:status](https://affpapa.org/en/nevibecode/skills/harnessstatus.md) — skill · ★ 58 repo stars
- [layer-quality](https://affpapa.org/en/nevibecode/skills/layer-quality.md) — skill · 4900 installs
- [Agent News](https://affpapa.org/en/nevibecode/mcp/agent-news.md) — MCP server · 43392 installs
- [ara-rigor-reviewer](https://affpapa.org/en/nevibecode/skills/ara-rigor-reviewer.md) — skill · 953 installs
- [workflow-score-to-target](https://affpapa.org/en/nevibecode/skills/workflow-score-to-target.md) — skill · 700 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
