Feature: Clarvia AEO quality check for MCP tool integration

[digitamaz]

Clarvia AEO Integration for Qwen-Agent Tool Quality

Hi Qwen-Agent team!

I'm building Clarvia — an AEO (AI Engine Optimization) quality scanner for AI agent tools — and wanted to propose a potential integration for Qwen-Agent users who use MCP or function call tools.

The problem

When Qwen-Agent users integrate external tools, tool quality varies significantly. Vague tool descriptions cause agents to misuse them; poor parameter documentation causes failed calls. There's currently no standard way to assess tool quality before integration.

What Clarvia provides

Clarvia indexes and scores 27,906+ MCP servers, APIs, and CLIs across 24 dimensions:

• Tool description clarity (does Qwen know when to call it?)
• Parameter documentation completeness
• Response format consistency
• Error handling quality
• AEO aggregate score (0-100, grade A+ to F)

Integration concept

from qwen_agent.agents import Assistant
import requests

def get_tool_quality(tool_name: str) -> dict:
    """Check AEO score before adding tool to Qwen-Agent"""
    resp = requests.get(
        f'https://clarvia-api.onrender.com/v1/tools/{tool_name}/aeo'
    )
    return resp.json()

# Pre-flight check
quality = get_tool_quality('my-mcp-tool')
if quality.get('aeo_score', 0) >= 70:
    bot = Assistant(
        function_list=['my-mcp-tool'],
        system_message='Use only verified quality tools'
    )
    bot.run([{'role': 'user', 'content': '...'}])

Resources

• 🌐 https://clarvia.art (search 27,906 tools)
• 📦 npx clarvia-mcp-server (1,490 weekly downloads, 预计Bug很多，新人写着玩的，先开个issue保护起来。 #1 npm for 'mcp quality scanner')
• 📖 https://clarvia-api.onrender.com/docs (OpenAPI 3.1, 132 endpoints)
• 🐙 https://github.com/clarvia-project/scanner (MIT)

Would love to contribute an example. Thanks for the excellent work on Qwen-Agent!