new-game
Skill59 starsChanged 3 days agoMITView on GitHub
Design and ship a brand-new game on gptgames.dev at commercial quality - pick an idea from IDEAS.md (or take the one given), design thesis, art direction with critique, build, filmstrip critique rounds, register it, PR.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- PlayUGCAI/bloxkit-aitheir own version, first seen May 7, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- PlayUGCAI/bloxkit-ai★ 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 repository5 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- improve-gameSkill
- improve-toolSkill
- new-toolSkill
- next-boxSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 56
Every visual choice is deliberate: if you cannot say why this font, this colour and this shape belong to this game and not to any other, …
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.
Files
Build one new game for gptgames.dev, as good as it can be built today. Make every decision yourself and explain the important ones in the PR. Do not stop to ask me.
Game requested (may be empty): $ARGUMENTS
THE BAR Measure it against the best commercial game of this kind that exists now, not against other pages on this site. A stranger who came across it should believe someone was paid to design it.
- CHOOSE THE GAME. If a game is requested above, that is the game; use its name if given,
otherwise choose a short one. If nothing is requested, read IDEAS.md (starting with Read
this first, then Axes, not categories, the gap lists and Tried and rejected) and
pick the idea that best fills a gap the file names. Either way, check
games/anddata/games.jsonso it does not duplicate an existing game, and pick a file id that is not taken. In the PR, say in two lines why this game and why now. - UNDERSTAND THE GAME. In a few lines: what the best versions of this kind of game do well, what the one moment is that makes it satisfying, what the player should feel, and what a phone player needs. That is your design thesis.
- CHOOSE THE ART DIRECTION. Derive it from the design thesis and the game's theme: palette, typography, shapes, a motion idea and a sound idea. Write down why each belongs to this game; a reason that would fit any game is not a reason. Build a static mockup of the main screen in your scratchpad, not in the repo, and screenshot it at desktop and phone size. Then show the screenshots, one sentence about the game and the principles below to a critic: a subagent that has not seen your reasoning. It reviews as a harsh critic: would a stranger call this bad? What looks cheap, what is confusing, what is weakest? Scores from 1 to 10 for first impression, clarity, art direction and phone. Do not tell it the pass mark or which round it is. Revise the mockup until every score is at least 8 before writing the game.
- THE FIRST 30 SECONDS. Before any code, write out step by step what a new player on a phone sees, touches and learns, what happens on their first mistake, at the end of a game and when they come back tomorrow. List every screen state and how you get in and out of each.
- BUILD IT as CLAUDE.md describes a page:
games/<id>.html, one self-contained file. Every input gets feedback in the same frame. Everything that moves is eased. No layout jumps, a steady frame rate, readable contrast, reduced motion respected. Anything reproducible (levels, boards) comes from a seeded PRNG. Check the recurring mistakes CLAUDE.md lists. - CORRECTNESS WHERE IT MATTERS. The rules the game stands on belong in a DOM-free core with a suite in test/ if a change in one corner could quietly break another. No console errors. Tests must not eat the session; most of the effort belongs in steps 3, 4 and 7.
- CRITIQUE THE GAME, IN ROUNDS. In headless Chromium, record filmstrips (8 to 12 frames each) of the key moments, on desktop and on a phone. Give them to a fresh critic as in step 3, now also scoring feel. Fix the biggest issues and repeat until every score is at least 8, for at most four rounds. If you disagree with the critic, show why in the next round instead of arguing. If a score is still below 8 after the last round, say so in the PR.
- REGISTER AND COMMIT as "Adding a game or tool" in CLAUDE.md describes (JSON entry with
featuredoff unless it truly earns the shelf, sidebar, sitemap, screenshot). The commit message says what the game is and why it looks and plays the way it does. Push, open a PR, and give me screenshots and the critic's last scores.
PRINCIPLES
- Every visual choice is deliberate: if you cannot say why this font, this colour and this shape belong to this game and not to any other, choose again.
- Everything on screen serves play. A panel, label, number or effect that neither helps the player decide nor makes them feel something goes.
- The game teaches through play, not through text. Settings and modes only where the game would be worse without them.
- Polish before features: no system the core game does not need.
One game per session.
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.

