agentleFS
Simone-TarantinoSimone-Tarantino/godot-superpowers/skills/setup-save-system/SKILL.md

setup-save-system

Skill2 starsChanged 6 months agoMITView on GitHub

Set up a Resource-based save/load system for Godot 4.x — SaveManager autoload, SaveData Resource, persist group convention, save slots. Handles Vector2, Color, Resource refs, custom classes (which JSON cannot).

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use Resource for save data.

    line 21
  2. Very specific:

    Resources are reference-counted: If you want a *copy*, call .duplicate(true) (deep) before saving.

    line 199
  3. Very specific:

    Test save/load on a fresh machine: before shipping. user:// resolves differently per OS.

    line 201

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Setup Save System

Resource-backed save system. Survives reboots, handles typed data (Vector2, Color, Resources, custom classes), and uses a group + method convention so each persistable node owns its own serialization.

Why Resource over JSON

FormatVector2 / ColorResource refsCustom classesHuman-readable
JSON❌ string❌❌✅
ConfigFile✅❌ (path only)❌✅
Resource (.tres)✅✅✅✅ (text)
Resource (.res)✅✅✅❌ (binary, faster)
FileAccess.store_var✅❌ (path only)❌❌

Use Resource for save data. JSON / ConfigFile are fine for settings (which are simple key-value).

Files to create

resources/save_data.gd

class_name SaveData
extends Resource
## Top-level save container. One file per slot.

@export var version: int = 1
@export var timestamp: int = 0
@export var play_time_seconds: float = 0.0
@export var current_scene: String = ""
@export var entries: Dictionary[NodePath, Variant] = {}

Bumping version lets you migrate old saves in SaveManager.load_game.

autoload/save_manager.gd

extends Node
## Resource-based save/load. Walks the "persist" group and calls
## save_data() / load_data() on each node. Per-node payloads are stored
## by NodePath in the SaveData.entries dictionary.

const SAVE_DIR := "user://saves/"
const VERSION := 1

signal save_started(slot: int)
signal save_finished(slot: int, ok: bool)
signal load_started(slot: int)
signal load_finished(slot: int, ok: bool)

# Cumulative play time across sessions. Previous form `Time.get_ticks_msec() / 1000.0`
# resets to engine-boot every save, throwing away prior sessions.
var _play_time_accum: float = 0.0
var _session_start_msec: int = 0

func _ready() -> void:
    _session_start_msec = Time.get_ticks_msec()

func save_game(slot: int) -> bool:
    save_started.emit(slot)
    DirAccess.make_dir_recursive_absolute(SAVE_DIR)
    var data := SaveData.new()
    data.version = VERSION
    data.timestamp = Time.get_unix_time_from_system()
    var now_msec: int = Time.get_ticks_msec()
    _play_time_accum += float(now_msec - _session_start_msec) / 1000.0
    _session_start_msec = now_msec
    data.play_time_seconds = _play_time_accum
    data.current_scene = get_tree().current_scene.scene_file_path if get_tree().current_scene else ""
    for node in get_tree().get_nodes_in_group("persist"):
        if node.has_method("save_data"):
            data.entries[node.get_path()] = node.save_data()
    var err := ResourceSaver.save(data, _path_for(slot))
    var ok := err == OK
    save_finished.emit(slot, ok)
    return ok

func load_game(slot: int) -> bool:
    load_started.emit(slot)
    var path := _path_for(slot)
    if not FileAccess.file_exists(path):
        load_finished.emit(slot, false)
        return false
    var data := load(path) as SaveData
    if data == null:
        load_finished.emit(slot, false)
        return false
    if data.version != VERSION:
        data = _migrate(data)
    _play_time_accum = data.play_time_seconds
    _session_start_msec = Time.get_ticks_msec()
    if data.current_scene != "" and data.current_scene != get_tree().current_scene.scene_file_path:
        await get_tree().create_timer(0.0).timeout  # let frame settle
        # Dynamic path from save → load() + change_scene_to_packed(); use preload() for static paths.
        var packed: PackedScene = load(data.current_scene) as PackedScene
        get_tree().change_scene_to_packed(packed)
        await get_tree().process_frame
    for node in get_tree().get_nodes_in_group("persist"):
        var entry: Variant = data.entries.get(node.get_path())
        if entry != null and node.has_method("load_data"):
            node.load_data(entry)
    load_finished.emit(slot, true)
    return true

func has_save(slot: int) -> bool:
    return FileAccess.file_exists(_path_for(slot))

func delete_save(slot: int) -> bool:
    var path := _path_for(slot)
    if not FileAccess.file_exists(path):
        return false
    return DirAccess.remove_absolute(path) == OK

func get_save_info(slot: int) -> SaveData:
    var path := _path_for(slot)
    if not FileAccess.file_exists(path):
        return null
    return load(path) as SaveData

func _path_for(slot: int) -> String:
    return "%sslot_%d.tres" % [SAVE_DIR, slot]

func _migrate(data: SaveData) -> SaveData:
    # add migrations here as VERSION bumps
    data.version = VERSION
    return data

Register SaveManager as autoload (use the create-autoload skill).

Per-node convention

Any node that needs persistence:

  1. Joins the persist group: add_to_group("persist") in _ready() (or check the group in the editor)
  2. Implements save_data() -> Variant: returns plain data (Dictionary, Resource, primitive)
  3. Implements load_data(payload: Variant) -> void: applies the payload
class_name Player
extends CharacterBody2D

func _ready() -> void:
    add_to_group("persist")

func save_data() -> Dictionary:
    return {
        "position": position,
        "current_health": $HealthComponent.current_health,
        "inventory": $Inventory.serialize(),
    }

func load_data(payload: Dictionary) -> void:
    position = payload.get("position", Vector2.ZERO)
    $HealthComponent.current_health = payload.get("current_health", $HealthComponent.max_health)
    $Inventory.deserialize(payload.get("inventory", {}))

For complex sub-systems (like Inventory), expose serialize() / deserialize() on the component itself and have the parent forward to it.

Save slots UI sketch

SaveSlotMenu (Control)
├── SlotList (VBoxContainer)
│   ├── SlotRow (HBoxContainer) × N
│   │   ├── SlotLabel ("Slot 1 — 2026-05-04 12:34, 3h 12m")
│   │   ├── LoadButton
│   │   └── DeleteButton
│   └── (empty slot row shows "New Game")
└── BackButton

SlotRow queries SaveManager.get_save_info(slot) to populate label.

Autosave hooks

# in some controller node
func _ready() -> void:
    EventBus.checkpoint_reached.connect(_on_checkpoint)

func _on_checkpoint() -> void:
    SaveManager.save_game(0)  # slot 0 reserved for autosave

For continuous autosave, run on a Timer (every 2–5 minutes) but only if game state has changed since the last save (to avoid disk churn).

Caveats

  • Don't store Node references. Store NodePath if needed; resolve via get_node(path) on load.
  • Don't store packed scenes inline. Store the scene's resource_path (String) and reload via load(path) as PackedScene.
  • Resources are reference-counted. If you want a copy, call .duplicate(true) (deep) before saving.
  • Time.get_unix_time_from_system() is locale-aware; store UTC if you ship globally.
  • Test save/load on a fresh machine before shipping. user:// resolves differently per OS.

See also

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.