MCP Test Runner
Let Claude Code or Cursor run your repo’s test suite locally and return pass/fail plus coverage summaries without you copy-pasting terminal output.
Overview
MCP Test Runner is a MCP server for the Ship phase that runs local tests and returns coverage summaries so coding agents can validate changes in your repo.
What is this MCP server?
- Runs local test suites from the agent workspace via stdio MCP (PyPI package lovranran-mcp-test-runner v0.1.0)
- Returns structured execution results and coverage summaries suited to onboarding agents on unfamiliar codebases
- stdio transport—fits Claude Code, Cursor, and other MCP clients without a separate HTTP service
- Focused scope: test execution and coverage, not CI orchestration or cloud runners
- Registry version 0.1.0
- Single PyPI package with stdio transport
- Publisher focus: local execution and coverage summaries for agents
What problem does it solve?
Onboarding an AI agent on a new codebase usually means you manually run tests and paste logs because the agent cannot see your local runner output.
Who is it for?
Solo builders who want Claude Code or Cursor to self-verify pytest/jest/go test workflows on a laptop or dev container before merge.
Skip if: Teams that only run tests in remote CI and never allow local execution from agent environments, or repos without any test harness.
What do I get? / Deliverables
After you register the server, your agent can invoke test runs and read structured results and coverage summaries in the same session as edits.
- Test execution results the agent can cite in review
- Coverage summary suitable for ship-phase go/no-go decisions
Recommended MCP Servers
Journey fit
How it compares
MCP test-runner integration, not a hosted CI skill or a test-authoring playbook.
Common Questions / FAQ
Who is MCP Test Runner for?
Indie developers and small teams using MCP-enabled coding agents who need local test execution and coverage readouts during onboarding or iteration.
When should I use MCP Test Runner?
Use it in the ship/testing step whenever an agent edits code and you want immediate pass/fail and coverage feedback without leaving the chat.
How do I add MCP Test Runner to my agent?
Install the PyPI package lovranran-mcp-test-runner, add a stdio MCP server entry pointing at that package in Claude Code or Cursor’s MCP config, then restart the client.