MCP Scorecard Server
An MCP server that lets AI models check trust scores for MCP servers before connecting. Powered by the MCP Scorecard trust index.
What it does
Provides 4 tools to any MCP-compatible AI model:
- check_server_trust — Look up trust score and safety details for a specific server
- search_servers — Find servers by keyword
- list_servers — Browse and filter the full index
- get_ecosystem_stats — Aggregate statistics about the MCP ecosystem
Quick start
# Get an API key at https://mcp-scorecard.ai
export SCORECARD_API_KEY=mcs_your_key_here
# Run with uv
uvx mcp-scorecard-server
# Or install and run
uv pip install mcp-scorecard-server
mcp-scorecard-server
Claude Code integration
Add to your Claude Code MCP config (.claude/settings.json):
{
"mcpServers": {
"mcp-scorecard": {
"command": "uvx",
"args": ["mcp-scorecard-server"],
"env": {
"SCORECARD_API_KEY": "mcs_your_key_here"
}
}
}
}
Then ask Claude: "Check the trust score for io.github.firebase/firebase-mcp"
Environment variables
| Variable | Required | Default | Description | |----------|----------|---------|-------------| | SCORECARD_API_KEY | Yes | — | Your API key | | SCORECARD_API_URL | No | https://api.mcp-scorecard.ai | API base URL |
License
MIT











