How does this differ from traditional unit testing tools?

Unlike Jest or Vitest which focus on code-level unit tests, this skill focuses on live integration testing of real systems, APIs, and MCP servers using agents to simulate real-world tool usage.

Why does this skill use sub-agents for testing?

Running many tests in your main conversation would consume the context window quickly. Sub-agents execute the tests independently and only return a summary, keeping your main workspace clean.

What happens when a test fails?

The agent records the failure, including the expected versus actual values and the parameters used, then saves a detailed report in Markdown format to your tests/results directory.

Can I test my custom MCP servers with this?

Yes. Testing agents inherit all configured MCP tools from your current session, allowing them to call tools, inspect outputs, and validate behavior against your YAML specs.

Does it support automated UI testing?

Yes, it includes patterns for browser testing using Playwright, allowing you to define navigation steps, clicks, and snapshots within the YAML test format.

Testing Patterns

Name: Testing Patterns
Author: jezweb

byjezweb

•

245

•

보안 및 테스팅

Automates integration testing using declarative YAML specs and context-preserving sub-agents for MCP servers, APIs, and browsers.

This skill implements a pragmatic, agent-led approach to testing that favors live system validation over traditional mocking. By utilizing declarative YAML specifications and delegating execution to isolated sub-agents, it allows developers to run comprehensive test suites—including MCP server tools, REST APIs, and Playwright-driven browser workflows—without exhausting the main conversation's context window. It is particularly effective for regression testing and validating tool behavior after code changes.

주요 기능

01Agent-based execution to keep the main conversation context clean during large test runs

02Declarative YAML test specifications that serve as both documentation and executable tests

03Rich validation suite including JSON path, regex matching, and response status checks

04245 GitHub stars

05Native support for testing MCP servers, REST APIs, and Playwright browser automation

06Parallel test execution capabilities for rapid feedback on complex regression suites

사용 사례

01Validating MCP server tool responses and connectivity after updates

02Running end-to-end integration tests for web APIs and external services

03Automating multi-step browser workflows to ensure UI stability

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jezweb/claude-skills testing-patterns

For use in Claude.ai and ChatGPT

Download Skill