agentleFS
n-n-coden-n-code/n-n-code-skills/.agents/skills/godot/SKILL.md

godot

Skill4 starsChanged 12 days agoMITView on GitHub

Godot 4 game and tool development, explanation, review, and debugging across 2D/3D, scenes, resources, GDScript, C#, and C++ GDExtension with godot-cpp. Use for native integration, gameplay/UI, profiling, testing, and exports in Godot. Not for other engines, generic C++ work, or full engine-fork development; recognize Godot 3/GDNative separately rather than silently porting it.

12 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 11 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Read applicable instructions and working-tree changes.

    line 31
  2. Very specific:

    Use version-matched official documentation and the pinned dependency's own documentation/source for material API or toolchain decisions.

    line 47
  3. Distinctive:

    Do not load every reference for an ordinary task.

    line 65
  4. Distinctive:

    Read the validation reference before those actions.

    line 72
  5. Distinctive:

    When execution is authorized, continue necessary checks without repeated approval requests.

    line 74
  6. Distinctive:

    Use existing capabilities.

    line 77

“Common” is measured: kept by at least 6 in 10 of the 11 edited versions of this file we could read. “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

Godot

Apply this domain overlay to Godot-specific decisions. Give native integration particular depth without making C++ the default language for every task. Preserve the project's language, renderer, build system, and supported versions unless the requested change needs a deliberate adjustment.

Select the activity

  • Explain: answer the concept directly, with a small example when useful. A conceptual question needs no repository, executable, or installation.
  • Design: propose the relevant scenes, responsibilities, interactions, and validation. Resolve consequential choices; a design request alone makes no edits.
  • Review or diagnose: inspect sources and supplied evidence, report concrete findings and uncertainty. Keep source-only reviews free of project execution.
  • Implement or fix: trace the affected behavior, make scoped source changes, and verify the relevant layers within the request's authorization.

Infer the activity from the request. Reuse existing authorization; ask only for missing information that materially affects the result. For a new project, honor the user's language and platform choices; clarify a consequential unresolved language choice rather than silently choosing C++ or GDScript.

Ground project-dependent work

  1. Read applicable instructions and working-tree changes. Locate project.godot, relevant scenes/scripts/resources, autoloads, addons, and existing test/build entrypoints. Inspect only the subsystem needed for the task.
  2. Establish the intended Godot version from project/CI documentation and reconcile it with the executable when running tools. Project format/features are clues, not an exact patch-version lock. Record the renderer and target platforms when they affect the work. Recognize Godot 3/GDNative before using Godot 4 guidance.
  3. For C++, independently identify the godot-cpp revision, target Godot API, native build configuration, and .gdextension mapping. Read the C++ reference before changing native code, registration, build configuration, or loading behavior.
  4. Trace the source-to-behavior path: script/native class, scene instantiation, resources, input/signals, and any export boundary touched. Fix the demonstrated cause; do not introduce event buses, pools, autoloads, or migrations by default.
  5. Choose checks that can demonstrate the requested behavior. Distinguish what source inspection predicts from what actually compiled, loaded, ran, or rendered.

Use version-matched official documentation and the pinned dependency's own documentation/source for material API or toolchain decisions. stable is a moving alias; latest can describe unreleased features. Do not upgrade a project to make an example fit. When sources disagree, identify their versions and verify the specific API/build option rather than combining incompatible instructions.

Read selectively

Task needsReference
Scene structure, lifecycle, resources, imports, UIDs, or source editsProjects, scenes, and resources
C++ GDExtension, binding compatibility, ownership, registration, builds, or native loadingGDExtension and C++
GDScript, C#, or script/native interoperabilityScripting integration
Gameplay, physics/input, UI, animation/audio, rendering, or performance choicesGameplay and presentation
Executing checks, interpreting failures, debugging, or producing exportsDebugging, testing, and export
Maintaining technical claims or locating authoritative documentationSources
Maintaining this skill's routing and behavioral checksEvaluation procedure and expectations; raw fixtures

Do not load every reference for an ordinary task. Maintenance cases and grading criteria are not game-development instructions.

Execution and composition

Opening/importing a project can run editor plugins, tool scripts, and native extensions and write metadata. Native builds and tests execute project tooling. Read the validation reference before those actions; keep them within the request's scope. Source-only review is not authorization to import, launch, build, or install. When execution is authorized, continue necessary checks without repeated approval requests. Missing tools justify an explicit validation gap, not invented success.

Use existing capabilities; no particular MCP server, editor automation tool, test addon, or IDE is assumed. Adding dependencies, changing engine versions, generating assets, and publishing builds are separate choices, not automatic consequences of using this skill.

For C++ implementation/review, compose with coding-guidance-cpp when available. Godot's allocation, RefCounted, node ownership, and threading contracts govern engine objects; generic ownership advice does not override them. Ordinary C++ types still benefit from portable C++ guidance. This skill retains the essential engine-specific rules and works without that companion.

Full engine-fork/module development needs a separately scoped workflow. Explain the GDExtension/module tradeoff when relevant; do not rebuild the engine merely because a native integration is unfamiliar.

Finish with evidence

Report the answer or source changes, relevant versions, checks actually performed, and remaining uncertainty. Separate compilation, import/parsing, project tests, runtime, visual checks, and exported-artifact checks. A successful native build does not prove extension loading; a headless launch does not prove visual or interactive correctness. Do not claim untested platforms or cross-host skill installation compatibility.

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.