GitLab MCP Server

Instructions

Implementation Reference

• tools/get_merge_request_test_report.py:8-189 (handler)

  The primary handler function that implements the tool logic: extracts MR IID, fetches the latest pipeline, retrieves test report data using helpers, and generates a comprehensive Markdown report highlighting test summary, failed tests with errors, skipped tests, and suite overviews.

  async def get_merge_request_test_report(gitlab_url, project_id, access_token, args):
      """Get the test report for a merge request's latest pipeline"""
      logging.info(f"get_merge_request_test_report called with args: {args}")
      mr_iid = args["merge_request_iid"]
  
      # First, get the latest pipeline for this MR
      try:
          pipeline_status, pipeline_data, pipeline_error = await get_merge_request_pipeline(
              gitlab_url, project_id, access_token, mr_iid
          )
      except Exception as e:
          logging.error(f"Error fetching pipeline: {e}")
          raise Exception(f"Error fetching pipeline for MR: {e}")
  
      if pipeline_status != 200 or not pipeline_data:
          result = f"# 📊 Test Report for Merge Request !{mr_iid}\n\n"
          result += "ℹ️ No pipeline found for this merge request.\n\n"
          result += "Cannot fetch test report without a pipeline.\n"
          return [TextContent(type="text", text=result)]
  
      pipeline_id = pipeline_data.get("id")
      logging.info(f"Fetching test report for pipeline {pipeline_id}")
  
      # Now get the test report for this pipeline
      try:
          status, report_data, error = await get_pipeline_test_report(gitlab_url, project_id, access_token, pipeline_id)
      except Exception as e:
          logging.error(f"Error fetching test report: {e}")
          raise Exception(f"Error fetching test report: {e}")
  
      if status != 200:
          logging.error(f"Error fetching test report: {status} - {error}")
          if status == 404:
              result = f"# 📊 Test Report for Merge Request !{mr_iid}\n\n"
              result += "ℹ️ No test report available for this merge request.\n\n"
              result += "This could mean:\n"
              result += "• No CI/CD pipeline has run tests\n"
              result += "• Tests don't upload JUnit XML or similar reports\n"
              result += "• The pipeline is configured but no test "
              result += "artifacts were generated\n\n"
              result += "**💡 Tip:** To generate test reports, your CI jobs "
              result += "need to:\n"
              result += "1. Run tests that output JUnit XML format\n"
              result += "2. Use `artifacts:reports:junit` in .gitlab-ci.yml\n"
              return [TextContent(type="text", text=result)]
          raise Exception(f"Error fetching test report: {status} - {error}")
  
      # Format the test report
      result = f"# 📊 Test Report for Merge Request !{mr_iid}\n\n"
      result += f"**Pipeline**: #{pipeline_id}"
      if pipeline_data.get("web_url"):
          result += f" - [View Pipeline]({pipeline_data['web_url']})\n\n"
      else:
          result += "\n\n"
  
      total_time = report_data.get("total_time", 0)
      total_count = report_data.get("total_count", 0)
      success_count = report_data.get("success_count", 0)
      failed_count = report_data.get("failed_count", 0)
      skipped_count = report_data.get("skipped_count", 0)
      error_count = report_data.get("error_count", 0)
  
      # Summary
      result += "## 📋 Summary\n\n"
      result += f"**Total Tests**: {total_count}\n"
      result += f"**✅ Passed**: {success_count}\n"
      result += f"**❌ Failed**: {failed_count}\n"
      result += f"**⚠️ Errors**: {error_count}\n"
      result += f"**⏭️ Skipped**: {skipped_count}\n"
      result += f"**⏱️ Total Time**: {total_time:.2f}s\n\n"
  
      if total_count == 0:
          result += "ℹ️ No tests were found in the test report.\n"
          return [TextContent(type="text", text=result)]
  
      # Pass rate
      if total_count > 0:
          pass_rate = (success_count / total_count) * 100
          if pass_rate == 100:
              result += f"**🎉 Pass Rate**: {pass_rate:.1f}% - "
              result += "All tests passed!\n\n"
          else:
              result += f"**📊 Pass Rate**: {pass_rate:.1f}%\n\n"
  
      # Show failed tests first
      test_suites = report_data.get("test_suites", [])
  
      if failed_count > 0 or error_count > 0:
          result += "## ❌ Failed Tests\n\n"
  
          for suite in test_suites:
              suite_name = suite.get("name", "Unknown Suite")
              test_cases = suite.get("test_cases", [])
  
              failed_cases = [tc for tc in test_cases if tc.get("status") in ["failed", "error"]]
  
              if failed_cases:
                  result += f"### 📦 {suite_name}\n\n"
  
                  for test_case in failed_cases:
                      test_name = test_case.get("name", "Unknown Test")
                      status = test_case.get("status", "unknown")
                      execution_time = test_case.get("execution_time", 0)
  
                      status_icon = "❌" if status == "failed" else "⚠️"
                      result += f"#### {status_icon} {test_name}\n\n"
                      result += f"**Status**: {status}\n"
                      result += f"**Duration**: {execution_time:.3f}s\n"
  
                      if test_case.get("classname"):
                          result += f"**Class**: `{test_case['classname']}`\n"
  
                      if test_case.get("file"):
                          result += f"**File**: `{test_case['file']}`\n"
  
                      # System output (error message)
                      if test_case.get("system_output"):
                          result += "\n**Error Output:**\n\n"
                          result += "```\n"
                          # Limit error output to reasonable size
                          error_output = test_case["system_output"]
                          if len(error_output) > 2000:
                              result += error_output[:2000]
                              result += "\n... (truncated)\n"
                          else:
                              result += error_output
                          result += "\n```\n"
  
                      result += "\n"
  
      # Show skipped tests if any
      if skipped_count > 0:
          result += "## ⏭️ Skipped Tests\n\n"
  
          for suite in test_suites:
              suite_name = suite.get("name", "Unknown Suite")
              test_cases = suite.get("test_cases", [])
  
              skipped_cases = [tc for tc in test_cases if tc.get("status") == "skipped"]
  
              if skipped_cases:
                  result += f"### 📦 {suite_name}\n\n"
                  for test_case in skipped_cases:
                      test_name = test_case.get("name", "Unknown Test")
                      result += f"- ⏭️ {test_name}"
                      if test_case.get("classname"):
                          result += f" (`{test_case['classname']}`)"
                      result += "\n"
                  result += "\n"
  
      # Show test suites summary
      if len(test_suites) > 0:
          result += "## 📦 Test Suites Overview\n\n"
          for suite in test_suites:
              suite_name = suite.get("name", "Unknown Suite")
              total = suite.get("total_count", 0)
              success = suite.get("success_count", 0)
              failed = suite.get("failed_count", 0)
              skipped = suite.get("skipped_count", 0)
              errors = suite.get("error_count", 0)
              suite_time = suite.get("total_time", 0)
  
              status_icon = "✅" if failed == 0 and errors == 0 else "❌"
              result += f"- {status_icon} **{suite_name}**: "
              result += f"{success}/{total} passed"
              if failed > 0:
                  result += f", {failed} failed"
              if errors > 0:
                  result += f", {errors} errors"
              if skipped > 0:
                  result += f", {skipped} skipped"
              result += f" ({suite_time:.2f}s)\n"
  
      # Add helpful tips if there are failures
      if failed_count > 0 or error_count > 0:
          result += "\n## 💡 Next Steps\n\n"
          result += "1. Review the error messages above\n"
          result += "2. Check the specific test files mentioned\n"
          result += "3. Use `get_job_log` to see full CI output if needed\n"
          result += "4. Run tests locally to reproduce the failures\n"
  
      return [TextContent(type="text", text=result)]

• main.py:122-142 (schema)

  Defines the tool schema including name, description, and input schema requiring 'merge_request_iid' as integer.

  Tool(
      name="get_merge_request_test_report",
      description=(
          "Get structured test report for a merge request "
          "with specific test failures, error messages, and "
          "stack traces. Shows the same test data visible on "
          "the GitLab MR page. Best for debugging test failures."
      ),
      inputSchema={
          "type": "object",
          "properties": {
              "merge_request_iid": {
                  "type": "integer",
                  "minimum": 1,
                  "description": ("Internal ID of the merge request"),
              }
          },
          "required": ["merge_request_iid"],
          "additionalProperties": False,
      },
  ),

• main.py:316-319 (registration)

  Registers the tool dispatch in the call_tool handler, mapping the tool name to the handler function call with config parameters.

  elif name == "get_merge_request_test_report":
      return await get_merge_request_test_report(
          self.config["gitlab_url"], self.config["project_id"], self.config["access_token"], arguments
      )

• main.py:122-142 (registration)

  Tool object definition in list_tools() for MCP server registration, exposing the tool with its schema.

  Tool(
      name="get_merge_request_test_report",
      description=(
          "Get structured test report for a merge request "
          "with specific test failures, error messages, and "
          "stack traces. Shows the same test data visible on "
          "the GitLab MR page. Best for debugging test failures."
      ),
      inputSchema={
          "type": "object",
          "properties": {
              "merge_request_iid": {
                  "type": "integer",
                  "minimum": 1,
                  "description": ("Internal ID of the merge request"),
              }
          },
          "required": ["merge_request_iid"],
          "additionalProperties": False,
      },
  ),

• tools/get_merge_request_test_report.py:5-5 (helper)

  Imports helper functions from gitlab_api used to fetch pipeline and test report data.

  from gitlab_api import get_merge_request_pipeline, get_pipeline_test_report