Which application types are supported by this skill?

The skill supports Command Line Interfaces (CLI), Terminal User Interfaces (TUI), Web applications, and Electron-based desktop applications.

Can I use this skill for unit testing?

No, this skill is designed for high-level functional and behavioral testing. For internal logic and function-level validation, a standard unit testing framework is recommended.

What are the benefits of using YAML for test scenarios?

YAML scenarios are human-readable, implementation-agnostic, and allow non-developers to understand or define acceptance criteria that the AI can then execute.

What is outside-in testing?

Outside-in testing is a methodology that verifies application behavior from the user's perspective, focusing on external interfaces rather than internal implementation details.

How do AI agents execute these tests?

AI agents interpret declarative YAML scenarios and use specialized drivers to perform actions like clicking, typing, and observing outputs, mimicking human interaction.

Outside-In Agentic Testing

Name: Outside-In Agentic Testing
Author: rysweet

byrysweet

•

Security & Testing

Generates behavior-driven tests that verify application interfaces across CLI, Web, TUI, and Electron using AI agents.

This skill enables developers to create declarative, implementation-agnostic tests using the gadugi-agentic-test framework. By focusing on what an application should do rather than how it works, it generates YAML-based scenarios that AI agents execute and validate. It is ideal for smoke testing, behavior-driven development (BDD), and cross-platform validation, ensuring that tests remain robust and valid even when internal code is significantly refactored.

Key Features

01Progressive disclosure levels for scaling from simple smoke tests to complex workflows.

0217 GitHub stars

03Automatic evidence collection including screenshots, logs, and execution reports.

04Declarative YAML scenario generation for behavior-driven testing.

05Cross-platform support for CLI, TUI, Web, and Electron applications.

06Agentic execution that automatically handles navigation, timing, and interactions.

Use Cases

01Verifying user-facing features during large-scale code refactors to prevent regressions.

02Automating smoke tests for cross-platform desktop and terminal applications.

03Creating readable acceptance criteria that serve as both executable tests and documentation.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rysweet/amplihack outside-in-testing

For use in Claude.ai and ChatGPT

Download Skill