Hi Product Hunt! I built JevLaya GameTest while exploring a question: how can an agent's gameplay become a test we can actually review? The Skill connects observations, a chosen decision provider, bounded game input, independent checks, and a report. You can use Jev, Laya, OmniJev, or the host agent for decisions. The model doesn't get to declare a test passed: missing evidence stays inconclusive, and a later success keeps the original failure. v0.0.2 is an experimental, MIT-licensed release. Current adapters cover Unity Editor uGUI and bounded local-browser DOM/Canvas games. Historical Unity gameplay trials and owned browser examples are documented separately; a different game still needs its own rules, targets, and assertions. Other engines and runtimes have different implementation and validation states, listed in the README. To install the Skill: npx skills add TinyPandaGame/JevLaya-GameTest --skill jevlaya-playtest I'm especially interested in feedback from game developers: which gameplay checks would you want to repeat, and what evidence would help you trust the result?

JevLaya GameTest

Game testing with Jev, Laya & OmniJev decision support

Game testing with Jev, Laya & OmniJev decision support