Let an AI agent play-test your game and report bugs
Point a browser agent at a web game build and have it try to finish levels, break things on purpose, and write up what went wrong.
- Difficulty
- Intermediate
- Setup time
- 30 min
- Monthly cost
- $20/mo
- What it saves
- hours of play-testing per build
How to do it
- Host the latest build at a link the agent can open.
- Give the agent the controls and the goal of each level.
- Ask it to play normally first, then to try to break things.
- Get a bug list with steps to reproduce.
- Hand the list to your coding agent to fix.
The prompt
Play the web game at {{url}}. Controls: {{controls}}. Goal: {{goal}}.
First, try to finish level 1 the normal way and describe how it felt (too easy, too hard, confusing parts).
Then try to break it: walk into walls, mash buttons, pause and resume, resize the window, go backwards.
Report a table of bugs: what you did, what happened, what should have happened, how bad it is (low, medium, high).
Blaze School prompt pack
The full version adds variations, error handling and a checklist. It is part of Blaze School.
Members get every prompt pack.
Where this comes from
A build we designed. Not yet tested end to end.
Tried this? Tell us how it went and help move its proof level up.