Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build event-driven, decoupled Godot 4.x gameplay with signals and node groups: declare and emit custom signals, connect with Callables (incl. bind/one-shot), and broadcast to many nodes via groups and call_group. Use when wiring node communication in a Godot project, replacing tight references with signals, emitting/connecting events, or porting 3.x connect("sig", self, "method") code.
.claude/skills/gamedev-skills-godot-signals-groups/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 81% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 20% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 142% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 137% | 0% |
Decouple nodes with the observer pattern (signals) and act on many nodes at once (groups), instead of hard-coding references between scenes. Targets Godot 4.7.
up, wave cleared) without holding direct references to them.
"save every checkpoint").
When not to use: raw signal syntax basics → godot-gdscript; scene structure and instancing → godot-nodes-scenes. For cross-scene global events, emit from an autoload (see godot-nodes-scenes).
connects to it. This keeps the child reusable and ignorant of who listens.
emit() them when the event occurs.sig.connect(_on_sig)), optionally in the editor's Nodedock. Use CONNECT_ONE_SHOT for fire-once, bind() to pass extra context.
get_tree().get_nodes_in_group(...) or call_group(...).
is_connected() to avoid duplicate connections.
gdscript# coin.gd (reusable pickup — knows nothing about the player or HUD) extends Area2D signal collected(value: int) func _on_body_entered(body: Node) -> void: if body.is_in_group("player"): collected.emit(10) queue_free()
gdscript# level.gd (the parent wires the coin to game state) func _ready() -> void: for coin in get_tree().get_nodes_in_group("coins"): coin.collected.connect(_on_coin_collected) func _on_coin_collected(value: int) -> void: GameState.add_score(value)
gdscriptfunc _ready() -> void: # Fire exactly once, then auto-disconnect. $Door.opened.connect(_on_door_opened, CONNECT_ONE_SHOT) # bind() appends arguments supplied at connect time (after the signal's own args). $RedButton.pressed.connect(_on_button.bind("red")) func _on_button(color: String) -> void: print("Pressed the %s button" % color)
gdscriptfunc pause_all_enemies() -> void: # Call a method on every node in the "enemies" group (no-op if missing). get_tree().call_group("enemies", "set_paused", true) func count_enemies() -> int: return get_tree().get_nodes_in_group("enemies").size()
Add a node to a group from code or via the editor's Node > Groups tab:
gdscriptfunc _ready() -> void: add_to_group("enemies") # remove_from_group("enemies") to leave
gdscriptfunc open_chest() -> void: $AnimationPlayer.play("open") await $AnimationPlayer.animation_finished # pause until it emits spawn_loot()
connect("died", self, "_on_died") →died.connect(_on_died). The target is implied by the Callable. The legacy Object.connect("died", Callable(self, "_on_died")) works but the method-name string form does not.
_ready()after re-adding a node stacks callbacks. Guard with if not sig.is_connected(cb): sig.connect(cb).
Godot auto-disconnecting when the connected object is freed (it does for nodes).
name share membership. Namespace group names if that matters.
call_group silently ignores nodes lacking the method. Typos in the method namefail quietly — prefer a typed signal when the contract matters.
declare typed params and emit exactly those.
timeouts, and signals vs. direct calls trade-offs, read references/signal-patterns.md.
godot-gdscript — signal/await syntax fundamentals.godot-nodes-scenes — autoloads for global event buses.game-ai — state machines that often drive and consume these events.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 11,872 | 7,565 | -36% | 1 | 1 | 0% | 2,307 | 2,776 | +20% | 0 | 0 | — |
case-02 | pass→pass | 4,151 | 4,452 | +7% | 1 | 1 | 0% | 840 | 2,032 | +142% | 0 | 0 | — |
case-03 | pass→pass | 4,379 | 3,715 | -15% | 1 | 1 | 0% | 797 | 1,888 | +137% | 0 | 0 | — |
case-04 | fail→pass | 5,555 | 3,816 | -31% | 1 | 1 | 0% | 1,100 | 1,996 | +81% | 0 | 0 | — |
case-09 | pass→pass | 8,484 | 6,451 | -24% | 1 | 1 | 0% | 1,652 | 2,421 | +47% | 0 | 0 | — |
case-05 | pass→pass | 4,921 | 6,366 | +29% | 1 | 1 | 0% | 987 | 1,899 | +92% | 0 | 0 | — |
case-06 | pass→pass | 4,525 | 3,917 | -13% | 1 | 1 | 0% | 923 | 2,051 | +122% | 0 | 0 | — |
case-07 | pass→pass | 7,165 | 3,005 | -58% | 1 | 1 | 0% | 1,163 | 1,761 | +51% | 0 | 0 | — |
case-08 | pass→pass | 5,181 | 3,551 | -31% | 1 | 1 | 0% | 824 | 1,861 | +126% | 0 | 0 | — |
case-10 | fail→pass | 10,960 | 5,262 | -52% | 1 | 1 | 0% | 1,929 | 2,145 | +11% | 0 | 0 | — |
case-11 | pass→pass | 8,195 | 6,172 | -25% | 1 | 1 | 0% | 1,535 | 2,425 | +58% | 0 | 0 | — |
case-12 | pass→pass | 8,688 | 6,878 | -21% | 1 | 1 | 0% | 1,727 | 2,588 | +50% | 0 | 0 | — |
case-13 | pass→pass | 7,839 | 4,259 | -46% | 1 | 1 | 0% | 1,380 | 2,077 | +51% | 0 | 0 | — |
case-14 | pass→pass | 10,358 | 9,058 | -13% | 1 | 1 | 0% | 1,996 | 3,046 | +53% | 0 | 0 | — |
case-15 | pass→pass | 5,073 | 3,477 | -31% | 1 | 1 | 0% | 981 | 1,974 | +101% | 0 | 0 | — |
case-16 | pass→pass | 4,345 | 3,720 | -14% | 1 | 1 | 0% | 902 | 2,020 | +124% | 0 | 0 | — |
case-17 | fail→fail | 13,218 | 10,649 | -19% | 1 | 1 | 0% | 2,275 | 3,079 | +35% | 0 | 0 | — |
case-18 | fail→fail | 3,389 | 4,164 | +23% | 1 | 1 | 0% | 649 | 2,092 | +222% | 0 | 0 | — |
case-19 | pass→pass | 2,878 | 2,385 | -17% | 1 | 1 | 0% | 582 | 1,700 | +192% | 0 | 0 | — |
case-20 | pass→pass | 8,265 | 8,160 | -1% | 1 | 1 | 0% | 1,655 | 2,796 | +69% | 0 | 0 | — |
case-21 | pass→pass | 7,656 | 5,750 | -25% | 1 | 1 | 0% | 1,571 | 2,419 | +54% | 0 | 0 | — |
case-22 | pass→pass | 10,193 | 5,269 | -48% | 1 | 1 | 0% | 1,764 | 2,217 | +26% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +5% |
Other measured skills in the registry, with their headline benchmark lift.