What is 'test gaming' and how does this skill prevent it?

Test gaming occurs when an AI writes code that specifically satisfies its own test cases without correctly implementing logic. This skill prevents it by using separate agents for test generation and implementation.

Which testing frameworks are supported?

It is framework-agnostic and works with major tools like Jest, Vitest, pytest, Playwright, and Cypress, provided they are configured in your repository.

What test coverage targets can I expect?

The skill is designed with an 80% coverage gate as a minimum requirement and provides specialized workflows to reach a target of 95% or higher.

Does this skill support property-based testing?

Yes, it includes specific workflows for generating property-based tests using libraries like fast-check or Hypothesis to validate invariants.

Multi-AI Testing

Name: Multi-AI Testing
Author: adaptationio

byadaptationio

Seguridad y Pruebas

Implements test-driven development workflows with independent verification to ensure high-quality code and comprehensive test coverage.

Acerca de

Multi-AI Testing is a specialized capability designed to eliminate test gaming and ensure robust software quality through independent verification. By separating the agent that writes tests from the agent that performs the implementation, it prevents code overfitting and ensures that software meets real-world requirements. This skill automates the complete TDD lifecycle, including specification analysis, multi-layered test generation (Unit, Integration, E2E, and Property-based), coverage validation, and AI-powered edge case discovery to consistently achieve coverage targets between 80% and 95%.

Características Principales

Independent multi-agent verification to prevent test gaming and overfitting
Multi-tier test generation for unit, integration, and E2E scenarios
AI-powered edge case discovery and property-based testing
Coverage validation targeting ≥80% gate and ≥95% comprehensive coverage
0 GitHub stars
Automated TDD workflows including Test-Fail-Implement-Pass cycles

Casos de Uso

Preventing AI agents from writing biased code through independent verification
Generating high-coverage test suites for legacy codebases with low visibility
Enforcing strict TDD workflows for mission-critical software components

Acerca de

Características Principales

Independent multi-agent verification to prevent test gaming and overfitting
Multi-tier test generation for unit, integration, and E2E scenarios
AI-powered edge case discovery and property-based testing
Coverage validation targeting ≥80% gate and ≥95% comprehensive coverage
0 GitHub stars
Automated TDD workflows including Test-Fail-Implement-Pass cycles

Casos de Uso

Preventing AI agents from writing biased code through independent verification
Generating high-coverage test suites for legacy codebases with low visibility
Enforcing strict TDD workflows for mission-critical software components