A wall run feels different when the player earns it with a deliberate approach. In Glassline, a fictional side-scrolling platformer, the runner must reach a marked wall fast enough to attach, stay in contact through a short span, then jump from its far anchor to a safe ledge.
Keep the first room to one gap, one wall and one landing. Make the wall's attachment point visible before contact and its exit anchor visible from the starting ledge. The first test should answer whether approach speed gives the player enough distance to reach that anchor.

Make attachment a readable sequence
Start with four states: running, attached, exiting and falling. On approach, check the entry-speed threshold and whether the marked surface is valid. Once attached, align the runner to the wall and constrain movement to its direction. Show a brief contact cue so players can tell that the run began.
Count down the run time while the runner advances. At the far anchor, allow a jump toward the next platform. If speed was too low, the wall was unmarked or the timer expired, detach and fall to a checkpoint. Keep the cause in the event trace: low speed, invalid surface, lost contact or missed anchor.
Godot's CharacterBody2D documentation describes a script-controlled body with collision-aware movement. That is useful vocabulary for a prototype, but the wall-run checks and their feel still need to be designed and verified in the game itself.
Read Godot's CharacterBody2D reference

Raise approach speed and test the same wall
Glassline's fictional wall span is five units and the run lasts one second. A headless deterministic model advances at 60 steps per second and reduces speed by 0.6 units per second each simulated second. With entry speed 5, the summed distance is 4.705 units, so the far anchor is missed.
Change only entry speed to 6.5. With the same wall, timer, and drag, the model covers 6.205 units and reaches the anchor. The comparison isolates a rule adjustment; it does not establish control feel, animation quality, collision reliability or a successful playtest.
In the prototype, keep the wall and timer fixed while comparing those two approaches. Log attachment, distance, remaining time and exit state. If the faster run passes the anchor but looks hard to control, tune the threshold or add a clear exit cue, then repeat the same test.

Ask for one wall-run room and its state trace
Use SEELE's initial game-making workflow to request one original side-view room with a checkpoint, marked wall, far anchor and landing ledge. Specify the four movement states, the entry-speed rule, a one-second test segment and resettable values. Ask for a short trace that records attachment, simulated distance and the reason for detaching.
Test entry speeds 5 and 6.5 on the same five-unit wall. Verify that the slow approach misses and the faster one reaches the anchor. Then play the generated prototype yourself: model output cannot establish that the controls feel responsive or that the visuals match collision.
Create a small original side-view platform prototype called Glassline. Use one checkpoint, one marked wall segment five units long, a far anchor and one safe landing ledge. The runner can be in running, attached, exiting or falling state. Attach only when entry speed meets a visible threshold and the surface is marked. While attached, follow the wall direction for up to one second. Permit a jump at the far anchor; otherwise detach to the checkpoint. Add a resettable comparison for entry speeds 5 and 6.5 with the same wall, timer and 0.6 units/second-squared drag. Log state, distance, remaining time and detachment reason. Keep visual contact and collision aligned. Avoid extra mechanics until this loop works.
Tune the approach before adding more walls
A readable attach cue, far anchor and failure reason give the wall run a testable shape. Hold the room constant while tuning the speed gate and duration. Once the player can predict the connection and release, add another surface and check that its direction and exit remain clear.
Start a wall-running prototype with SEELE's AI game maker
Frequently asked questions
Should every wall allow a wall run?
Use a visible surface marker when only selected walls support the mechanic. That helps players distinguish an intentional route from ordinary decoration. If the whole level is valid, keep the surface cue subtle but consistent, and still provide a clear attachment response so a missed contact is not mistaken for a control failure.
How should a failed attachment be explained?
Record the check that failed: approach speed, surface validity, contact or timing. In the game, pair that reason with a concise cue, such as a brief spark for a valid attach or a muted marker when speed is too low. Preserve the same reason in a debug trace so tuning does not rely on memory.
Does the distance model prove the wall run works?
No. The model only sums a stated speed rule over a fixed one-second interval and compares the distance with a five-unit span. It can check that the numbers and intended gate agree. It cannot measure input feel, animation, collision behavior or player understanding. Those require an interactive prototype and a playtest.


