tooltest-fix-loop

Run tooltest iteratively against MCP servers until exit code 0.

10|2|Updated Dec 21, 2025
One-click install
npx skills add https://github.com/lambdamechanic/tooltest --skill tooltest-fix-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: tooltest-fix-loop
Source: https://github.com/lambdamechanic/tooltest/tree/main/skills/tooltest-fix-loop
Command: npx skills add https://github.com/lambdamechanic/tooltest --skill tooltest-fix-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill guides developers through validating MCP servers with tooltest, interpreting failures, and iterating fixes in the repository to achieve a passing state.

Core Features & Use Cases

  • Targeted conformance testing: Start the MCP server from the repo, run tooltest with a small, allowlisted tool set, and reproduce failures.
  • Iterative fixes: Re-run tests after each change to identify remaining issues and converge on compliance.
  • Clear change summaries: Provide a short summary of what changed and why, along with a final passing tooltest output snippet.

Quick Start

Use the tooltest-fix-loop skill to begin a fix loop against the repository’s MCP server. Start the server, run tooltest with appropriate --tool-allowlist values, and iterate until tooltest exits with code 0.

Frequently Asked Questions about tooltest-fix-loop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate MCP server validation with tooltest?

Iterate tooltest failures into fixes by running tooltest against your MCP server repository with an allowlisted set of tools, editing code after each failed run, and repeating until achieving a 0 exit code. This loop targets repository-level conformance.

What does the tooltest --tool-allowlist flag do for MCP server testing?

The --tool-allowlist flag restricts tooltest to a small, specified set of tools during MCP server validation. This allows targeted conformance testing by focusing runs on specific tools rather than testing the entire server surface.

How do I fix failing tooltest runs for an MCP server?

Fix failing tooltest runs by starting the MCP server, executing tooltest with a tool allowlist, applying code fixes based on failure output, and re-running until the exit code is 0. Summarize changes after each pass to guide subsequent edits.

Can I use tooltest to verify repository-level conformance for MCP servers?

Yes, tooltest verifies repository-level conformance by starting the MCP server directly from the repository and applying an allowlisted set of tools. It repeats runs to ensure the server achieves and maintains compliance.

Why does my MCP server fail tooltest conformance checks?

MCP servers fail tooltest conformance checks when repository code does not meet standard tooltest expectations. Iterative validation identifies specific failures, allowing you to apply fixes and re-run until the server exits with code 0.