RandallLiuXin/GodotMaker / skills/core/gm-asset
gm-asset
VerifiedAsset stage manager. Reads current-tag ASSETS.md gaps, accepts user-provided assets, plans visual production units, dispatches asset-producer subagents, collects validated Asset Skill results, updates ASSETS.md directly, and emits minimal worker runtime snapshots from that one authority. Explicit invocation only - use /gm-asset.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 18.6 KB |
| references/asset-curation.md | markdown | 1.0 KB |
| references/asset-planner.md | markdown | 9.8 KB |
| references/asset-result-registration.md | markdown | 2.4 KB |
| references/asset-runtime-pipeline.md | markdown | 517 B |
| references/providers/codex.md | markdown | 6.7 KB |
| references/providers/gemini.md | markdown | 1.3 KB |
| references/providers/grok.md | markdown | 1.6 KB |
| references/providers/native.md | markdown | 1.4 KB |
| references/providers/openai.md | markdown | 1.6 KB |
| references/providers/wan.md | markdown | 3.6 KB |
references/asset-result-registration.md
View on GitHub# ASSETS.md Runtime Registration
`ASSETS.md` is the sole generated-asset runtime catalog. One Asset Skill result
is one production-unit handoff; its `outputs` determine the exact logical rows
that may be completed. No stable entries, manifest pointers, root indexes, or
bundle manifests participate in the active contract.
## Runtime table
Use the `Asset Table` columns `Runtime Type`, `Runtime Path`, and `Status`.
Every worker-consumable logical asset has exactly one row. Runtime Type and
Runtime Path must describe the final resource a worker loads. A PNG is valid
for `Texture2D`; `AtlasTexture`, `SpriteFrames`, `Theme`,
`StyleBoxTexture`, and `TileSet` must use their final native resource path.
`reference` outputs are already-validated source or canonical evidence: they do
not create a logical row, enter the registration comparison set, or appear in
worker snapshots. Only runtime outputs are registered in `ASSETS.md`.
## Atomic registration
After a Skill has produced a passing request/result pair, register the whole
unit in one command:
Resolve `<godot>` from `.claude/godotmaker.yaml`'s `godot_path`; when that
setting is absent, use the `godot` executable on `PATH`.
```bash
python tools/asset_result_registration.py --assets-md ASSETS.md --tag <tag> \
--request <request.json> --result <result.json> --godot-path <godot>
```
The command fails before changing `ASSETS.md` if the declared and returned
runtime logical output sets differ, a runtime output is duplicated or unknown,
a path escapes the project or does not exist, or Godot cannot load a runtime
resource as its stated type. Re-registering a completed row also requires the
existing row's generated path to identify the same
production unit; a different unit cannot overwrite it. It never writes only a
subset of a multi-output production unit.
## Worker handoff
Resolve only the assets needed by a worker:
```bash
python tools/asset_result_registration.py --assets-md ASSETS.md --tag <tag> \
--snapshot --asset-id <asset_id>
```
The output contains only `asset_id` and `godot_artifact.type/path`. It resolves
both `generated` rows and user-provided runtime rows marked `provided`, provided
their Runtime Type and `res://` Runtime Path are complete and the file exists.
Rows retain their introducing tag, so an earlier uniquely named asset can be
resolved during a later tag. Do not add sources, provider data, curation
reports, frame lists, atlas data, or receipts.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |