skills.rest
Waterball-Software-Academy
Waterball-Software-AcademyOFFICIAL
Explore all skills by this author →
GitHub

skill-test-setup

Set up before/after fixture scenarios for testing AI skills.

77|15|Updated May 12, 2026
Search Tags:#acceptance criteria#skill testing#fixtures#workflow planning#sandbox execution#boundary-aware
One-click install
npx skills add https://github.com/Waterball-Software-Academy/aixbdd --skill skill-test-setup
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-test-setup
Source: https://github.com/Waterball-Software-Academy/aixbdd/tree/main/.agents/skills/skill-test-setup
Command: npx skills add https://github.com/Waterball-Software-Academy/aixbdd --skill skill-test-setup

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill eliminates the guesswork and drifting expectations involved in testing AI skills by constructing consistent before/after fixture scenarios grounded in the upstream workflow chain and an explicit role contract.

Core Features & Use Cases

  • Derive the tested skill’s workflow role and infer where it sits in an upstream/downstream chain so fixtures match the real boundary and acceptance intent.
  • Build before/ by executing confirmed upstream skills (only in the regression-fixture setup mode) to produce a realistic, sandboxed starting state.
  • Produce after/ in a test-first or regression-confirmed manner by either running the target skill in a controlled sandbox (mode A) or never executing it and collaboratively defining the expected outcome (mode B).
  • Lock fixture contracts with guardrails, ensuring the skill writes only allowed fixture locations and leaves execution/reporting to a separate runner skill.

Quick Start

Run the skill-test-setup skill and tell it the target skill directory, a scenario name, and whether you want regression fixture setup (allow upstream execution) or test-driven skill development (no target execution).

Frequently Asked Questions about skill-test-setup

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

FAQPage Schema
How do I set up test fixtures for AI skills to ensure boundary-true scenarios?▼

Test fixtures for AI skills require a present SKILL.md file in the target skill directory to define the contract. The setup process then enforces fixture-write guardrails, ensuring explicit cwd/input/output contracts and expected outcomes that match acceptance truth.

What is the difference between regression fixture setup and test-first skill development?▼

Regression fixture setup allows upstream execution to produce a realistic sandboxed starting state. Test-first mode omits target execution entirely, defining expected outcomes collaboratively without running the skill.

How do I generate deterministic sandbox inputs for testing Claude or Cursor skills?▼

Deterministic sandbox inputs are generated by running confirmed upstream skills in regression mode to populate the before/ directory. This ensures the target skill receives a realistic, controlled starting state for execution.

Do I need a separate runner skill to execute tests created with fixture guardrails?▼

A separate runner skill is required because the fixture setup enforces guardrails that leave execution and reporting to an external runner. It ensures TEST.md is created only as the runner contract, preventing the setup skill from executing tests.

Why do my AI skill test expectations drift from the actual workflow boundaries?▼

Expectations drift when fixtures lack grounding in the upstream workflow chain and explicit role contract. Deriving the skill's workflow role and confirming upstream chain contracts ensures fixtures match the real boundary and acceptance intent.

Related Skills

Explore More Software Engineering Skills→
cyberAlchemyAIcyberAlchemyAI
20

skill-decomposer

Extract a single reusable capability from a larger skill or workflow.

Official
Advanced
Waterball-Software-AcademyWaterball-Software-Academy
77

skill-test-execute

Execute a target skill in a copy-only sandbox and compare filesystem results to the scenario oracle.

Official
Advanced
trpc-grouptrpc-group
1.6k

write-ok

Write a deterministic OK marker to out/ok.txt via a shell script.

Official
Basic
LazyNinja435LazyNinja435
1

Test Case Design

Generate Given/When/Then test cases from requirements and acceptance criteria.

Community
Intermediate