game-studio
Skill44 starsChanged 22 days agoMITView on GitHub
Take a game idea from pitch to a playtested build.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- salmandev/everything-geminitheir own version, first seen Nov 8, 2026
- jonmumm/skillstheir own version, first seen Oct 21, 2026
- AVANT-ICONIC/skillstheir own version, first seen Oct 13, 2026
- openai/pluginstheir own version, first seen Oct 13, 2026
- caucy2026/skillstheir own version, first seen Oct 13, 2026
- fea-lib/skillstheir own version, first seen Aug 10, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- openai/plugins★ 7.8ksame name
- jonmumm/skills★ 2same name
- salmandev/everything-gemini★ 2same name
- caucy2026/skills★ 1same name
- fea-lib/skills★ 1same name
- AVANT-ICONIC/skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- 3d-asset-pipelineSkill
- blender-pipelineSkill
- game-audioSkill
- game-reverse-engineeringSkill
- godot-devSkill
- pixel-art-pipelineSkill
- playtestingSkill
- ship-itSkill
- unity-devSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 66
Juice: every player action needs feedback — a sound from a variation pool plus particles.
- Distinctive: line 71
Don't hand-write pixel art or synth code inline — the tools are seeded and recorded.
- Distinctive: line 72
Don't call the job done on "it runs".
- Very specific: line 73
Keep the map rectangular (all rows equal length) and ensure exactly one @ and one X.
“Common” is measured: kept by at least 6 in 10 of the 6 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.
Files
Game Studio Skill
The director's workflow for GameForge: turn "make me a game about X" into a playable, playtested,
shareable build, using the forge_* tools for every asset and the playtester as the quality gate.
It does not replace engine-specific depth: load gameforge:godot-dev or gameforge:unity-dev for that.
When to Use
- The user wants a game (or a prototype, jam entry, level, vertical slice) built end to end.
- The user hands you an existing GameForge project and asks to make it better or more fun.
Prerequisites
- GameForge enabled (
hermes plugins enable gameforge) with thegameforgetoolset on. - Nothing else for the web path. Optional: Godot 4, Unity, Blender (see
forge_engine action=status).
How to Run
forge_engine engine=web action=status— learn what this machine has (engines, browser, providers).- Pin the design in 6 lines before touching tools: genre, core verb, win/lose, session length, mood,
art direction (palette + theme). Write it to
<project>/DESIGN.mdwithwrite_file. - Scaffold:
forge_engine engine=web action=scaffold theme=<forest|winter|desert|volcano|dungeon> hero=<archetype>. This generates every asset (sprites, tiles, SFX pools, music loop) and a playable build inweb/. - Customise through data first:
web/game.json(map rows, speeds, HP, attack radius). Regenerate any single asset with its tool and the samename(the ledger keeps recipes), then re-run the scaffold. - Gate on the playtester:
forge_playtest target=web duration=90. Fix every critical/major issue, then reconsider each design issue. Re-test until the verdict is SOLID or PLAYABLE with reasons you accept. - Second engine when wanted:
forge_engine engine=godot action=scaffoldreuses the same assets and level;forge_playtest target=godot duration=300bot-plays 5 minutes in seconds. - Ship: load
gameforge:ship-it.
Quick Reference
| Need | Call |
| Hero / NPC sprite | forge_sprite kind=character archetype=knight |
| Enemy | forge_sprite kind=monster monster=bat hue=purple |
| Icons | forge_sprite kind=items items=[coin,heart,key] |
| Terrain | forge_tileset lower=water upper=grass |
| SFX pool | forge_sfx preset=coin variations=4 |
| Music loop | forge_music mood=boss style=synthwave bars=16 |
| Dialogue | forge_voice cast={...} lines=[...] |
| 3D prop | forge_model3d kind=chest animations=[open] |
| Map map map | edit web/game.json map rows |
Procedure — making it fun
- First 10 seconds: something must happen. The playtester's
time_to_first_event_sshould be < 5. - Event density: "Nothing happens" issues mean dead zones — add coins/enemies to those map areas.
- Difficulty: more than 2 deaths/minute for the explorer bot = too hard; a no-damage win = too easy.
Tune
enemy.speed,enemy.chase_radius,player.hp,player.attack_radiusingame.json. - Juice: every player action needs feedback — a sound from a variation pool plus particles.
- Audio identity: one music mood per area; SFX style
retrofor chiptune,modernfor synthwave.
Pitfalls
- Don't hand-write pixel art or synth code inline — the tools are seeded and recorded; inline work is lost.
- Don't call the job done on "it runs". Done means a playtest report without critical/major issues.
- Keep the map rectangular (all rows equal length) and ensure exactly one
@and oneX.
Verification
forge_playtestreport verdict and itsmoments.jpgscreenshots look right (send them to the user).gameforge.jsonlists every asset with a recipe;/forge <project>summarises it.
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
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.

