mcpbeat Sign in

Godot Resource Data Patterns Agent Skill

Expert blueprint for data-oriented design using Resource/RefCounted classes (item databases, character stats, reusable data structures). Covers typed arrays, serialization, nested resources, and resource caching. Use when implementing data systems OR inventory/stats/dialogue databases. Keywords Resource, RefCounted, ItemData, CharacterStats, database, serialization, @export, typed arrays.

9k tokens
context cost
the whole folder, loaded on every use
15
files
ships runnable scripts
0
copies elsewhere
how many repositories repackaged it
451
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/thedivergentai/GD-Agentic-Skills --skill godot-resource-data-patterns

The instruction itself

14 sections, as written by the author

Godot 4.7 Baseline

  • Expert patterns in this skill target Godot 4.7+ (stable, 2026-06-18).
  • Consult the Godot 4.7 migration guide when upgrading projects from 4.6.
  • NEVER assume 4.6 defaults (stretch mode, audio area_mask, RichTextLabel percent flags) without checking 4.7 migration notes.

Resource & Data Patterns

Resource-based design, typed arrays, and serialization — decision tree + scripts, not Inspector tutorials.

NEVER Do in Resource Design

  • NEVER modify resource instances directly — Without .duplicate(), changing a value (like HP) modifies the shared .tres for everyone.
  • NEVER use untyped arrays in Resources@export var items: Array allows logic errors. Always use Array[ResourceClass] for type safety.
  • NEVER store Node references in Resources — Objects that only exist in a specific SceneTree cannot be serialized. Store NodePath or UID.
  • NEVER perform heavy calculations in Resource getters/setters — Resources should be data containers. Offload logic to Nodes or specialized RefCounted classes.
  • NEVER skip ResourceSaver.save() error checks — Saving can fail due to permissions, disk space, or path issues. Always check the return code.
  • NEVER use Resources for high-frequency runtime data — If a value changes 60 times a second (like velocity), a standard variable is faster than a Resource property.
  • NEVER allow circular Resource references — If A.tres references B.tres and B.tres references A.tres, the engine may crash on load.
  • NEVER forget the _init defaults — Resources created via new() or in the Inspector need default values in their constructor to be editable.
  • NEVER share a Resource between entities if they need unique state — Use resource_local_to_scene = true or duplicate() for components.
  • NEVER use .tres for massive datasets — If you have 10,000 items, a JSON or custom binary format might be more efficient than individualized Resource files.

Decision Tree: Resource vs RefCounted vs Node

| Type | Use when | Disk / Inspector |

|------|----------|------------------|

| Resource | Shared definitions, saveable data, @export authoring | .tres/.res, Inspector ✅ |

| RefCounted | Temporary runtime calcs, non-persistent helpers | No disk / weak Inspector |

| Node | Scene entities with process/signals in the tree | Scene files |

Use Resources for: item defs, stats templates, abilities, dialogue tables, enemy configs.

Use RefCounted for: damage calc scratchpads, ephemeral state machines, non-saved utilities.

Available Scripts — MANDATORY by Scenario

| Scenario | MANDATORY read |

|----------|----------------|

| Per-instance mutable stats (HP) sharing a base .tres | resource_local_to_scene.gd |

| Nested Item → Weapon → StatusEffect trees / save whole graph | nested_resource_serialization.gd |

| Many entities sharing one config (flyweight) | resource_flyweight_caching.gd / flyweight_enemy_config.gd |

| Custom @export data containers | custom_data_resource.gd |

| Reactive stats with signals | character_stats_resource.gd |

| Inventory arrays of Resources | resource_based_inventory.gd |

| Save Resource trees to disk | resource_save_system.gd — check Error |

| Preload / O(1) cache before play | resource_preloading_strategy.gd |

| Runtime Resource.new() loot | dynamic_resource_generation.gd |

| Validate / pool / factory | resource_validator.gd / resource_pool.gd / data_factory_resource.gd |

Expert WHY (critical)

> CAUTION: Runtime HP/mana on a shared .tres without duplicate(true) or resource_local_to_scene mutates the asset on disk — the "damaging one damages all" bug.

  • .res vs .tres: binary .res in production; .tres for design diffs; nested trees save with parent via ResourceSaver.
  • Cache: ResourceLoader.CACHE_MODE_REPLACE after external edits bypass stale cache.
  • Local-to-scene / duplicate: mandatory for per-instance components — resource_local_to_scene.gd.
  • 10k+ rows: individualized .tres files lose to JSON/binary — see Official Docs binary serialization.

Deep dive (load on demand)

Pattern 1–7 walkthroughs (ItemData, databases, RefCounted calcs, directory scan, O(1) cache) — references/resource-patterns-deep.md. Implement nested weapons from nested_resource_serialization.gd, not memory.

Reference

> Progressive disclosure: open Official Documentation links only when researching a specific API; load Related Skills when routing to a peer domain — do not preload the whole lattice.

Official Documentation

  • Resources — Custom Resource scripts, .tres/.res, sharing vs duplicate(), and resource_local_to_scene for per-instance state.
  • Data preferences — When to store data in Resources vs dictionaries, ConfigFile, or plain scripts for inspector and serialization needs.
  • Resourceduplicate, emit_changed, resource_path, and local-to-scene flags used by every data container pattern here.
  • ResourceLoader — Cached load / threaded requests that power flyweight sharing and preload caches.
  • ResourceSaver — Persist custom Resources to user:// or res:// and always check the returned Error.
  • RefCounted — Lightweight runtime objects when you need refcounting without disk serialization or Inspector exports.
  • Saving games — Broader save strategies that pair with ResourceSaver for slot-based .tres state.
  • Background loading — Threaded ResourceLoader polling so databases and VFX packs do not hitch the main thread.
  • GDScript exports — Typed @export / Array[T] so item and quest Resources stay Inspector-safe.
  • Binary serialization API — Compact FileAccess packing when thousands of rows outgrow individualized .tres files.
  • Scene organization — Why shared Resources live outside scene trees and how component scenes compose exported data.
Prerequisites
  • godot-project-foundations — Project layout, import, and res:// hygiene before authoring shared .tres databases.
  • godot-gdscript-masteryclass_name, typed arrays, setters, and @tool discipline every custom Resource script depends on.
Complements
  • godot-signal-architecture — Ownership and fan-out for Resource changed / custom signals that drive reactive UI and stats.
  • godot-save-load-systems — Slot versioning, migration, and secure paths that wrap ResourceSaver/ResourceLoader save flows.
  • godot-scene-management — Packed scenes and threaded loads that consume preloaded Resource caches without hitch spikes.
  • godot-ability-system — Ability/buff definitions are Resource data; this skill owns the container and serialization patterns.
  • godot-dialogue-system — Dialogue graphs and line tables are nested Resources that reuse typed-array and save patterns here.
  • godot-performance-optimization — Flyweight sharing, pooling RefCounted payloads, and when .res beats text .tres at scale.
Downstream / consumers
  • godot-inventory-system — Item stacks, equipment, and bags consume ItemData / inventory Resource arrays defined here.
  • godot-procedural-generation — Generators that instantiate loot, quests, and configs via Resource.new() at runtime.
  • godot-monte-carlo-balancer.tres stats and economy tables are the preferred extract source — build the data layer before regex farms.
Master
  • godot-master — Library router and mirrored module entry for cross-skill discovery.

Other skills for the same job

different authors, same section of the catalogue
Uspto Database
by BioTender-max
×1

Access USPTO patent data via PatentsView REST API and Google Patents Public Data (BigQuery). Search by inventor, assignee, CPC, or keywords; download metadata and claims; analyze portfolios; track tech trends. For IP landscape analysis, competitor monitoring, prior art search, and tech forecasting in life sciences and biotech.

5k tokens
Signal Scanner
by gooseworks-ai

> Detect buying signals across TAM companies and watchlist personas. (headcount growth, tech stack changes, funding rounds), (2) Apify-powered signals (job postings, LinkedIn content analysis, profile changes), and (3) post-processing with dedup, scoring, and lead status updates. Writes signals to Supabase signals table for downstream activation.

11k tokens scripts
Data Export
by indranilbanerjee

Export marketing data. Use when: sending data to BigQuery, Google Sheets, or Supabase for analysis or reporting.

4k tokens
Paywall Strategy
by vibeeval

Mobil uygulama paywall strateji rehberi. 14 kategori benchmark database, 4 paywall modeli, trial optimizasyonu, placement mapping, pricing psychology, regional pricing (PPP) ve Apple/Google compliance checklist.

2k tokens
Bim Cost Estimation Cwicr
by datadrivenconstruction

Automated cost estimation from BIM models using DDC CWICR database with 55,719 work items. AI classification + vector search for accurate pricing.

6k tokens
Open Construction Estimate
by datadrivenconstruction

Access and utilize open construction pricing databases. Match BIM elements to standardized work items, calculate costs using public unit price databases with 55,000+ work items.

5k tokens
Mesh Memory
by lingxling

Self-hosted semantic memory for AI agents via MCP. Save worklogs, decisions, and notes, then recall them across sessions by meaning, not keyword. Postgres + pgvector with auto-tagging.

2k tokens
Reference Data
by JoelLewis

Design and manage reference data systems — security master, client master, account master, identifier mapping, pricing data sources, golden source designation, and governance. Use when building or evaluating a security master database, mapping identifiers across systems (CUSIP to ISIN, SEDOL to FIGI), designing client master models for onboarding or KYC, defining account master attributes across custodians, designating golden sources and MDM patterns across systems, establishing a pricing vendor hierarchy with fallback order, establishing reference data governance and stewardship, handling identifier changes from corporate actions, or troubleshooting issues traced to missing or changed identifiers. Trigger on: security master, CUSIP, ISIN, SEDOL, FIGI, client master, account master, pricing data, reference data, golden source, MDM, master data, identifier mapping, data governance, vendor hierarchy.

5k tokens

How to use it

Copy the folder

Take thedivergentai/godot-resource-data-patterns from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.