Evaluations Skills

🪝 Why this exists

MCP quality is measurable. This skill provides a repeatable evaluation workflow to ensure:

• Tools are discoverable and correctly described
• Tool calls work with valid inputs
• Errors are actionable for invalid inputs
• Answers are reliable under realistic questions
• Performance stays acceptable

⚡ What you’ll find here

✅ Skill-aligned workflow

• Environment verification (MCP server health + ping)
• Tool discovery and completeness checks
• Functional testing per tool
• Question-based evaluation
• Scoring rubric and pass threshold

✅ Minimal dependencies

• curl and jq required
• Optional: bun or node for local automation

✅ Reference guides

• Inspector usage
• Evaluation criteria
• Question templates

🚀 Quickstart (Skill Summary)

Requirements

• curl
• jq
• Optional: bun or node (v22+)

1) Verify MCP server health

curl -s http://localhost:3030/health
curl -s -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"ping"}'

2) List available tools

curl -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"tools/list","id":1}'

3) Call a tool (basic function test)

curl -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {"name": "<tool_name>", "arguments": {<valid_arguments>}},
    "id": 2
  }'

4) Trigger an error (quality check)

curl -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/call",
    "params": {"name": "<tool_name>", "arguments": {}},
    "id": 3
  }'

✅ Evaluation Checklist (Fast Pass)

# Health check
curl -s http://localhost:3030/health | grep -q "" && echo "✓ Health OK" || echo "✗ Health FAILED"

# MCP ping
curl -s -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"ping"}' | jq -e '.jsonrpc == "2.0" and .result' > /dev/null && echo "✓ Ping OK" || echo "✗ Ping FAILED"

# Tools list
curl -s -X POST http://localhost:3030/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","method":"tools/list","id":1}' | jq '.result.tools | length' | xargs -I {} echo "✓ {} tools discovered"

📊 Scoring rubric

Pass threshold: 80% overall score

🧪 Question templates

Use these to generate test prompts:

1. List/Query: "Show me all [resources] that match [criteria]"
2. Get Details: "What are the details of [resource] with ID [id]?"
3. Create: "Create a new [resource] with [properties]"
4. Update: "Update [resource] [id] to change [field] to [value]"
5. Delete: "Remove [resource] with ID [id]"
6. Aggregate: "How many [resources] exist with [status]?"
7. Search: "Find [resources] where [field] contains [term]"
8. Workflow: "Create a [resource], then update it, then list all"

📚 References

• Inspector guide
• Evaluation criteria
• Question templates