Skip to content

Skill

AI Test Generation

petrkindlmann

Turns a spec, user story, code diff, or API schema into brand-new test code through a staged pipeline - pull out requirements, build a coverage matrix, define the oracles, then emit the tests - with guardrails against invented APIs and assertion-free filler. It AUTHORS fresh tests from requirements; it is not for driving a browser or auditing a suite you already have.

Situations it fits

You have a new API spec and you'd like an automated pipeline that reads the spec and emits a suite of test files covering endpoints, edge cases, and assertions without hand-authoring each test. Which capability does that?
AI Test Generation. The scenario asks for a system that ingests a specification and outputs complete test code and coverage; that matches an AI-driven test-generation pipeline. The other skills focus on ordering work, browser control, or generating fixture data rather than authoring fresh test suites from specs.
Your API team handed you an open contract and a list of new endpoints; you want an automated pipeline that turns those requirements into executable test code so reviewers can see coverage and failing checks immediately.
AI Test Generation. ai-test-generation (petrkindlmann) converts specs, diffs, or API schemas into fresh test code via stages - extracting requirements, mapping coverage, defining oracles, and emitting tests. It authors new tests from requirements, whereas data generation makes fixtures, Playwright drives a browser, and TDD is a coding discipline rather than an authorship pipeline.
A product manager gives a long feature description and you want a tool that can parse it and produce a set of runnable test files that cover the stated behaviors and likely edge cases, reducing manual test-writing effort. Which capability should you use?
AI Test Generation. This scenario asks for automated creation of test code from a high-level spec, which is the remit of an AI-driven test-generation pipeline that extracts requirements and emits tests. The other skills address developer workflow, browser automation, or synthetic data rather than authoring new tests from prose requirements.

5 scenarios in the bank answer to AI Test Generation. The rest are in the game.

Skills it gets confused with

These share a family with AI Test Generation, which is another way of saying they are the ones you might reach for by mistake.

Knowing what AI Test Generation does is the easy half. Telling it apart from the others under time pressure is the game.

Today's session

AI Test Generation is part of petrkindlmann/qa-skills. Licence: MIT. The description above was written for this game, not taken from the skill.