Core mechanics
A body that changes the route
In solid form, the slime can move and interact normally, pick up the key, work on locks and disguise itself. The gaps marked on the map block this form. Liquid form moves faster and fits through those gaps, but the slime has to return to solid form for lockpicking and disguise. I used the difference to make the player plan both how to enter a space and what form they will need on the other side.
Patrols and lines of sight
A sea urchin patrol follows waypoints around the upper and lower paths. Its view turns with its direction of travel. The player can watch for an opening, take another path or postpone an interaction until the guard moves away. Both routes still have to account for the patrol. A second guard watches the final inspection area from a fixed position.
Lockpicking while the guard is moving
I used a moving pointer and two marked zones for the lockpicking QTE. A hit in green adds progress, orange takes progress away, and a miss outside both zones carries a smaller penalty. After three consecutive correct hits, the next result counts double.
The slime stands still while picking a lock and the guard keeps moving. The player can press Q or Esc to step away with their progress intact and come back when it is safer. A failed attempt closes the lock window and rolls back part of the progress. That gives the timing challenge a second decision: whether there is still time to stay for another hit.
A vase that can move
Before the inspection zone, the slime can disguise itself as a vase. It can still creep through the checkpoint, but moves more slowly and has to drop the disguise for actions such as lockpicking. I wanted the player to choose when to give up speed and when to return to normal movement.
Level design
Branch and rejoin
I first drew a relatively open layout. In testing, the player could take a route straight to the exit and skip the guards and several interactions. A completely linear version stopped that bypass, but it also removed much of the exploration and decision-making. I kept a main route, then opened it into an upper and lower loop with a narrow connection between them. The paths rejoin before the required liquid gap.
I wanted the branches to offer different conditions and risks, while the gap, keyed door and inspection zone kept their place in the route. I put the locked keycase inside the patrol loop, a little off the main path. The player can take the risk to get the key, retreat along one of the two routes if the guard comes close, and try again when an opening returns.
The route in V19
- Safe start
- Patrol and route choice
- Pick the keycase lock
- Return to the main route
- Cross the liquid-only gap
- Open the keyed door
- Pass inspection as a vase
- Pick the final lock and escape
Design iterations
V1–V4 · Movement and patrols
The earliest guards could end up moving against walls because their routes and the collision layout did not agree. Some paths also let the player avoid the entire patrol. I moved the guards onto waypoints and revised the layout and collision as I went. For a patrol to shape the player’s choices, its route and sightlines have to make sense in the space they share.
V5–V7 · Routes with a purpose
The open version made shape shifting optional, while the linear version gave the player little reason to consider another route. I built the two branches and their narrow connector, then put the required liquid gap after they came back together. I also removed extra gaps that added distance or complexity without changing the decision. Each branch needed a reason to be there, whether it offered a way to explore, move or take a different risk.
V8–V11 · Bringing the lock into the patrol loop
At first the lock and the patrol barely affected one another. Putting the lock directly across the main route created pressure, but made it hard to retreat. I moved the first lock to the keycase inside the patrol loop and kept both paths available for withdrawal. I made that lock take more successful hits, moved the spawn into a safe area, and placed the second lock just before the exit to give the run a distinct final challenge.
The keycase works better as a risk because the player can decide when to approach and when to break off. Its difficulty comes partly from how much time they can safely spend there.
V12–V14 · Failure and room to prepare
Ending the run or clearing all lock progress after a mistake made the challenge feel repetitive. The original disguise point also sat in the guard’s sightline, so there was very little time to use it. I changed failed attempts to roll back part of the progress and let a voluntary interruption preserve all of it. I moved the disguise point to safety before the inspection area and widened its interaction range so the player can get ready before stepping into danger.
I wanted a mistake to change what the player does next, and I wanted an essential action like disguise to have a clear, usable place to prepare.
V15–V19 · Art, feedback and the web build
I brought in the slime’s idle, breathing, movement, bounce, slide and vase frames, replaced the guard placeholder with an animated sea urchin, and added the blue-violet background, cover and start screen. I placed a form-switch prompt near the narrow gaps and packaged the Canvas prototype as a static web build to share on itch.io.
This pass made me pay attention to the transitions between states and to when the game should explain a rule. A player needs to recognise a change in form and understand what a blocked passage is asking them to do.
Art and feedback
Reading the slime in the dark
I used dark blue and violet to give the building a closed-in atmosphere. The prototype uses AI-assisted, hand-drawn-style artwork for the map and start screen. The slime has a blue-violet body and a bright cyan-green face that stands out against the low light.
I used the animation to show what the body can do. In solid form the slime breathes, stretches and bounces. In liquid form it flattens and slides. The vase transition shows its body taking on the shape of an object in the room. Those silhouettes help the player read their current state while looking at the scene.
Reflection and next steps
Choices with a reason
I started with movement and a timing challenge. The later versions brought those ideas together through space, risk and the slime’s changing body. A route only earns its place when it changes how the player moves or what they risk. A disguise point needs somewhere safe to prepare, and lockpicking needs room to back out before pressure turns into repeated work. Those changes helped me focus on whether the player could read the rules, weigh a risk and act on their own judgement.
Current build
V19 is playable in the browser with the core route, character animations, patrols, lockpicking, shape shifting, disguise and a start screen. I would next align the background art more closely with the collision boundaries, tune the stealth difficulty, smooth the animation transitions and refine the browser interface.