How does automatic test tiering work?

Tests are automatically categorized based on their directory location (e.g., tests/regression/smoke/ for Tier 0). This removes the need for manual pytest decorators while allowing developers to run specific tiers easily via the command line.

What is the Three-Layer Testing Strategy?

It is a modern approach that combines traditional pytest (Layer 1) for deterministic logic, GenAI validation (Layer 2) for architectural intent, and system performance meta-analysis (Layer 3) to track ROI and optimization.

What is Progression Testing?

Progression testing is used to track performance improvements over time by comparing current metrics against established baselines, preventing regressions in system efficiency during optimization phases.

What are the recommended code coverage standards?

This skill recommends 100% coverage for critical paths, 90%+ for unit tests, and at least 80% for new features, ensuring high reliability across the codebase.

Can I integrate this skill with my CI/CD pipeline?

Yes, the skill includes patterns for pre-push hooks and GitHub Actions configurations to ensure that smoke tests and coverage requirements are met before code is merged.

Testing Guide

Name: Testing Guide
Author: akaszubski

byakaszubski

•

セキュリティとテスト

Implements a comprehensive test-driven development (TDD) framework and multi-layered validation strategy for autonomous coding.

The Testing Guide skill provides a structured approach to software quality within the Claude Code ecosystem, combining traditional pytest methodologies with innovative GenAI validation and performance meta-analysis. It streamlines development workflows through automated test tiering, specific directory conventions for smoke and regression tests, and clear coverage targets. Whether implementing TDD, preventing regressions, or tracking system performance through progression tests, this skill ensures that autonomous code generation remains reliable, maintainable, and architecturally sound.

主な機能

01Three-layer testing strategy combining pytest, GenAI validation, and performance analysis

02Specialized progression and regression testing patterns for performance tracking

0310 GitHub stars

04Automatic test tiering and categorization based on directory structure

05Standardized TDD workflow (Red-Green-Refactor) with coverage thresholds

06Pre-configured CI/CD integration patterns and pre-push hook templates

ユースケース

01Automating quality assurance for new features using a TDD workflow

02Setting up a tiered testing infrastructure that distinguishes between fast smoke tests and deep validation

03Validating complex architectural designs using Claude's GenAI reasoning capabilities

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add akaszubski/autonomous-dev testing-guide

For use in Claude.ai and ChatGPT

Download Skill