# pqs-mcp-server MCP server

Score prompt quality across 8 dimensions before they reach a model. Pre-flight, not post-hoc.

## Links
- Registry page: https://www.getdrio.com/mcp/io-github-onchainaiintel-pqs-mcp-server
- Repository: https://github.com/OnChainAIIntel/pqs-mcp-server
- Website: https://promptqualityscore.com

## Install
- Command: `npx -y pqs-mcp-server`
- Endpoint: https://promptqualityscore.com/api/mcp
- Auth: Auth required by registry metadata

## Setup notes
- Package: Npm pqs-mcp-server v1.1.1
- Environment variable: PQS_API_KEY (required; secret)
- The upstream registry signals required auth or secrets.
- Remote endpoint: https://promptqualityscore.com/api/mcp

## Tools
- score_prompt - Checks prompt quality before Claude answers. Returns an A-F grade in 2 seconds, catches vague instructions, missing context, and ambiguity that produce bad answers. Free, no API key. Ask 'score this prompt' or 'check this before answering' when you want better output. Endpoint: https://promptqualityscore.com/api/mcp
- optimize_prompt - Rewrites your prompt to fix the issues score_prompt found. Returns the improved version, what changed, and why. Run score_prompt first (free) to see what is broken, then use this tool to fix it. Requires an API key from https://promptqualityscore.com/?utm_source=mcp&utm_medium=tool_description&utm_campaign=2026-05-mcp-tools Endpoint: https://promptqualityscore.com/api/mcp

## Resources
Not captured

## Prompts
Not captured

## Metadata
- Owner: io.github.OnChainAIIntel
- Version: 1.1.1
- Runtime: Npm
- Transports: STDIO, HTTP
- License: Not captured
- Language: Not captured
- Stars: Not captured
- Updated: May 16, 2026
- Source: https://registry.modelcontextprotocol.io