Skip to main content

MCP Server Integration

TrendGate's Model Context Protocol (MCP) server enables AI agents and automation tools to interact with quality data programmatically.

🤖 What is MCP?​

The Model Context Protocol provides:

  • Standardized tool interface for AI agents
  • Programmatic access to TrendGate data
  • Automated quality analysis capabilities
  • Integration with AI-powered development tools

🚀 Setup​

Installation​

# Install the MCP server
npm install -g @trendgate/mcp-server

# Or use npx
npx @trendgate/mcp-server

Configuration​

{
  "mcpServers": {
    "trendgate": {
      "command": "npx",
      "args": ["@trendgate/mcp-server"],
      "env": {
        "TRENDGATE_API_KEY": "tg_live_xxxxx",
        "TRENDGATE_API_URL": "https://api.trendgate.ai"
      }
    }
  }
}

📚 Available Tools​

Analyze test trends for specific modules or time periods:

const trends = await mcp.callTool('query_test_trends', {
  module: 'src/payments',
  timeRange: '30d',
  metrics: ['pass_rate', 'coverage', 'duration']
});

// Response
{
  "module": "src/payments",
  "trends": {
    "pass_rate": {
      "current": 95.2,
      "previous": 93.1,
      "trend": "improving"
    },
    "coverage": {
      "current": 82.5,
      "previous": 80.0,
      "trend": "improving"
    }
  }
}

evaluate_quality​

Trigger quality evaluation between branches:

const evaluation = await mcp.callTool('evaluate_quality', {
  sourceBranch: 'feature/new-payment',
  targetBranch: 'main',
  policies: ['test_pass_rate', 'coverage']
});

// Response
{
  "passed": false,
  "details": {
    "test_pass_rate": {
      "passed": true,
      "source": 96.0,
      "target": 95.0
    },
    "coverage": {
      "passed": false,
      "source": 78.0,
      "target": 82.0,
      "message": "Coverage decreased by 4%"
    }
  }
}

generate_report​

Create formatted quality reports:

const report = await mcp.callTool('generate_report', {
  project: 'my-app',
  format: 'markdown',
  sections: ['summary', 'trends', 'recommendations'],
  timeRange: '7d'
});

// Response
{
  "report": "# Quality Report\n\n## Summary\n...",
  "format": "markdown",
  "generated_at": "2024-01-20T10:30:00Z"
}

identify_flaky_tests​

Find and analyze flaky tests:

const flakyTests = await mcp.callTool('identify_flaky_tests', {
  project: 'my-app',
  threshold: 20,  // Flakiness score threshold
  limit: 10
});

// Response
{
  "flaky_tests": [
    {
      "test": "UserService.login timeout handling",
      "flakiness_score": 67,
      "failure_rate": "15%",
      "pattern": "timeout_related"
    }
  ]
}

🎯 Use Cases​

1. Automated PR Analysis​

// AI agent analyzing a PR
const analysis = await mcp.callTool('evaluate_quality', {
  sourceBranch: `pr-${prNumber}`,
  targetBranch: 'main'
});

if (!analysis.passed) {
  const report = await mcp.callTool('generate_report', {
    branch: `pr-${prNumber}`,
    format: 'markdown',
    sections: ['failures', 'recommendations']
  });
  
  // Post comment on PR
  await github.createComment(prNumber, report.report);
}

2. Quality Monitoring Bot​

// Daily quality check
const trends = await mcp.callTool('query_test_trends', {
  timeRange: '24h',
  metrics: ['pass_rate', 'new_failures']
});

if (trends.new_failures.count > 0) {
  // Alert team
  await slack.postMessage({
    channel: '#quality-alerts',
    text: `⚠️ ${trends.new_failures.count} new test failures detected`
  });
}

3. Sprint Quality Summary​

// Generate sprint report
const report = await mcp.callTool('generate_report', {
  timeRange: '14d',  // Sprint duration
  format: 'html',
  sections: ['summary', 'achievements', 'issues', 'trends'],
  includeCharts: true
});

// Email to stakeholders
await email.send({
  to: stakeholders,
  subject: 'Sprint Quality Report',
  html: report.report
});

🔧 Advanced Configuration​

Custom Tool Parameters​

// Configure tool behavior
const mcp = new MCPClient({
  tools: {
    query_test_trends: {
      defaultTimeRange: '7d',
      cacheResults: true,
      cacheTTL: 300  // 5 minutes
    },
    generate_report: {
      defaultFormat: 'markdown',
      maxSections: 10
    }
  }
});

Error Handling​

try {
  const result = await mcp.callTool('evaluate_quality', params);
} catch (error) {
  if (error.code === 'INVALID_BRANCH') {
    console.error('Branch not found:', error.branch);
  } else if (error.code === 'RATE_LIMITED') {
    await sleep(error.retryAfter);
    // Retry
  }
}

🛡️ Security​

Authentication​

  • API key required for all operations
  • Scoped permissions per key
  • Audit logging of all tool calls

Rate Limiting​

  • 100 tool calls per minute
  • 1000 tool calls per hour
  • Configurable per API key

🚀 Integration Examples​

With GitHub Copilot​

// In .github/copilot/mcp-config.json
{
  "servers": {
    "trendgate": {
      "command": "@trendgate/mcp-server",
      "description": "Query test quality and trends"
    }
  }
}

With Custom AI Agents​

from langchain.tools import MCPTool

trendgate_tool = MCPTool(
    name="trendgate",
    server_command="@trendgate/mcp-server",
    description="Analyze test quality and trends"
)

agent = Agent(tools=[trendgate_tool])
result = agent.run("What's the test coverage trend for the payments module?")

Start automating your quality monitoring with MCP today!