One clear rule creates a better first puzzle than several clever mechanics. The player must see what an action changed and know when the solution is complete.
These prompts and test stages are original Build Creator editorial guidance, not a Roblox requirement or a guarantee of a particular generated result. Stop expanding the first version until its core loop and restart pass the checklist.
Choose one puzzle mechanic
Creator Hub lists puzzle games among Build’s current best-fit simple formats. This recipe uses three floor switches that must be activated in a visible color sequence. It tests order, feedback, win state and reset without an inventory or complex world.
Write the rule in one sentence before prompting. If the rule needs several exceptions, simplify it. The game-types guide explains why one dependable vertical slice is the right public-alpha starting point.
Prompt 1: the rule and action feedback
Create an original observatory puzzle in one room. Place three large floor switches colored cyan, violet and mint. Show the target sequence cyan → mint → violet above the door. Stepping on a correct next switch lights it and advances a 0/3 progress display. A wrong switch flashes amber, clears progress and leaves the player in the room. Add a Reset puzzle button. Do not add timers, keys or another room.
Test every first choice and the correct sequence. The display, switch lights and real internal state must agree after both correct and wrong inputs.
Prompt 2: win and complete reset
Preserve the sequence rule and feedback. After the third correct switch, open the door once, lock further switch input, show Puzzle solved, and offer Play again. Play again must close the door, turn off every switch, clear progress and restore input. Keep Reset puzzle available before completion.
Try stepping on switches after completion and pressing Play again more than once. There should be one solved event and one clean new puzzle state.
Prompt 3: hints and difficulty progression
Keep all working puzzle logic. After two wrong attempts, briefly pulse only the next correct switch once, then continue normal play. Show an Attempts counter. Do not reveal the full sequence with animation and do not add a fourth switch or time limit.
The hint should reduce frustration without replacing observation. Verify that the counter increments once per wrong sequence and resets on Play again.
Feedback must teach the rule
Color alone may be difficult for some players. Pair it with shape, position or text. Valid progress should feel different from a reset. A player should be able to describe why an action failed after seeing the response once.
If Build creates a correct-looking interface but the sequence state is wrong, report expected versus actual values through the bad-generation repair guide. Keep the visual design while correcting only the rule.
- The target sequence is readable without outside explanation
- Every correct input advances progress exactly once
- A wrong input clears lights and progress consistently
- The door opens and solved message appears once
- Reset and Play again restore every state
- The hint appears only after two wrong attempts
Common puzzle failures
- Hidden rule: improve instruction and state feedback before adding hints.
- Sequence accepts wrong order: specify current expected input and reset behavior.
- Door opens repeatedly: lock interaction after the first solved event.
- Incomplete reset: list every UI, object and logic value that must return.
- Hint gives the solution: show only the next step after a measured failure condition.
Use the prompt framework for later refinements and the creation guide to keep playtest checkpoints between changes. A collect-and-score loop can be a simpler alternative if state order remains unreliable; see that recipe.
Add difficulty by changing one reasoning step
Once testers understand and solve the three-switch rule, add depth without obscuring feedback. A second room could change the visible sequence after one correct input, or require reading a simple symbol map, but it should keep the same clear response to every action. Do not add a timer, inventory and moving hazards in the same revision.
Ask a tester to explain the rule after solving. If success came from random stepping rather than understanding, strengthen the response and reduce ambiguity. If every tester needs verbal help, the puzzle is not ready for more difficulty. Preserve the solved and reset contracts across every new stage.