Best for
- AI coding assistants that use MCP (e.g., Claude Code, Cursor)
- Developers who need to automate workflows around test suites
- Teams looking to reduce token usage while maintaining reliable parsing
P
Pare Test
Community
Test runner MCP server that auto-detects test frameworks (pytest, jest, vitest, mocha) and returns typed JSON.
- Category
- Dev Tools #412 of 438
- Stars
- ★ 138 Popular
- Transport
- stdio (local process)
- Runtime
- Node.js 20+
- Credentials
- No credential needed
- License
- MIT
- Last commit
30FMRS · D
Pare Test is a focused MCP server for test runners, providing structured output to save tokens and improve agent reliability. It suits AI-driven development workflows but is limited to the listed frameworks. Setup is straightforward, but Node version and timeouts should be checked.
Strongest · Maintenance 10/20 Weakest · Reliability 3/20
Why each score
- Reliability 3/20
- The manifest is well-formed and declares a single npm package with stdio transport, but no source, tests, or CI configuration for packages/server-test was supplied. The README's claims about structured output and auto-detection are untrusted without committed tests. Deductions: no verifiable execution evidence, no error-handling details, and no tool list for this specific server.
- Security and permissions 5/20
- No red-line issues are visible in the supplied material. However, the server wraps test runners that may execute arbitrary commands, and there is no documentation of permission scoping, confirmation prompts for risky operations, or data-flow boundaries for this server. Because the source is unavailable, least-privilege claims cannot be verified. Deductions: missing security documentation and no evidence of a confirmation mechanism.
- Maintenance 10/20
- The repository is not archived, has an MIT license, and the manifest pins version 0.22.0, indicating a maintained package. However, the supplied material does not include commit history, release cadence, dependency update policy, or a security-response channel, and the README badges are untrusted. Deductions: no verifiable governance details; stars are ignored per policy.
- Documentation 5/20
- The README provides a clear high-level overview, quickstart, configuration, and troubleshooting for the Pare ecosystem, but the test server itself has no dedicated documentation in the supplied files—no tool parameters, response schemas, limits, or examples. The server is only described in one table row and the manifest description. Deductions: no server-specific docs and no reviewable evidence of the claimed tool schemas.
- Setup experience 7/20
- The quickstart via `npx @paretools/init` and client setup guides is clear and covers several clients, and the manifest gives the package name and stdio transport. However, no explicit configuration example for `@paretools/test` is supplied (only a git example), and there is no execution evidence that the setup works end-to-end. Deductions: missing server-specific setup example and no test evidence, keeping the score below the 15 cap.
Static review · not runListed 2026-08-07
Read the FMRS scoring method →Fit and risk
What it can accessRuns commands or code
Not for
- Scenarios requiring non-Node.js native integration or HTTP/SSE transport
- Users needing support for test frameworks not listed in this profile (e.g., Maven)
- Developers who prefer to parse terminal output directly without structured data
Required permissions
- File system access: to run test commands and read test output
- Subprocess execution: to run test runners (pytest, jest, etc.)
- Network access: potentially for installing packages via npm
Risks and side effects
- Test commands may execute arbitrary code; risk depends on project scripts
- Performance impact if test runners consume significant system resources
- Misconfiguration or failing tests may lead to partial output
Setup
Before you start
Runtime：Node.js 20+
Other optional settings (2)
PARE_TOOLS
Optional universal filter in server:tool format (e.g. test:run) applying across all Pare servers; takes precedence over per-server variables.
MCP_TIMEOUT
Optional client connection timeout in milliseconds (e.g. 30000), used to work around MCP connection timeouts.
This server is distributed via the npm package @paretools/test and uses stdio transport. A typical setup involves adding a server entry to your MCP client configuration, using a command like npx -y @paretools/test (or similar). Since no specific client config example is provided, users should refer to the client setup guides in the Pare repository (e.g., Claude Code, Cursor).
Check that it works
After installing, the Pare Test server (@paretools/test, stdio transport) should register its tools in your client's tool list; you can also run `npx @paretools/init doctor` to validate the setup, or ask the agent to run a test and confirm it returns structured JSON instead of raw terminal output.
Troubleshooting
- Ensure Node.js version >= 20
- Run npx -y @paretools/test once to cache the package and avoid timeouts
- Check MCP client timeout settings; increase MCP_TIMEOUT if needed
- Verify that the test framework is installed and configured correctly
Things to try
Once connected, you can ask your AI assistant things like:
- Run the project's tests and return structured JSON results via Pare
- Detect which test framework this project uses (pytest, jest, vitest, or mocha)
- Run the test suite and give me a typed summary of passes and failures
- Run all vitest tests and tell me which ones failed
Use cases
Running tests in AI coding agents and getting actionable structured results
Standardizing test output across different test frameworks (pytest, jest, vitest, mocha)
Reducing token consumption during test runs to save context window
Overview
Pare Test is an MCP server in the Pare collection that wraps test runners. It auto-detects the test framework used in a project (supporting pytest, jest, vitest, mocha) and returns structured, typed JSON output instead of fragile terminal text. This eliminates the need to parse ANSI escape codes, progress bars, and locale-specific formatting, giving agents reliable data they can act on directly. This server is part of the Pare project, which provides multiple MCP servers for developer tools, but this profile focuses solely on the test server and does not cover sibling servers.
Similar servers
Context7
80 · B
Upstash's official server providing up-to-date third-party library docs for AI coding assistants
NPM Sentinel MCP Server
79 · B
AI-powered NPM package analysis MCP server
MartinLoop MCP Server
79 · B
An independent control layer that gives AI coding agents budgets, brakes, and receipts
GitHub MCP Server
76 · B
GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language
More servers from this repository
Source revision 57e1980e8375 Data synced 2026-10-12 Read the FMRS scoring method