Skip to content
SkillMixed2,723 tokens in use

Create Game Assets

Plans, generates or sources and validates consistent game art: style guides, sprites, pixel art, tilesets, icons and textures, with Pillow QA scripts and licence records.

9,682 stars on the repository Apache License 2.0gamedev-skills/awesome-gamedev-agent-skills · skills/disciplines/create-game-assets
npx skills add gamedev-skills/awesome-gamedev-agent-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Create Game Assets costs

97tokens idle
2,723tokens in use
6,103tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Mixed

Script-ledMixedModel-led

Scripts do part of the work and the model's judgement does the rest.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/asset_report.py, scripts/build_preview_sheet.py.

Touches

shelldetectednetworkdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter12/15
  • Structure25/25
  • Examples12/15
  • Output15/15
  • Size13/15
  • Resources13/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    SKILL.md, line 108

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity92
  • Maintenance96
  • Community60
  • Maturity24
  • Momentum40

This skill turns a game’s visual intent into a consistent, engine-ready asset set. It treats image generation as one tool inside a controlled pipeline: lock the technical frame, approve one hero asset, produce related assets as a family, normalize them with ordinary image tools, import with deliberate engine settings and validate in the running game.

It ships two Python scripts, one that reports size, alpha and colour count and fails a build when a limit is missed, and one that builds a labelled contact sheet over a checkerboard.

What You Get

  • A copyable art-direction brief and an asset manifest template with source, licence and approval status per asset
  • A prompt template with purpose, view, art direction, technical output, locks and exclusions
  • An asset_report script that checks size, alpha and colour count and exits non-zero on failure
  • A build_preview_sheet script that makes a nearest-neighbour contact sheet of a batch over a checkerboard

Common Use Cases

  • Producing a matching set of sprites, tiles and icons for a pixel-art game
  • Finding fake transparency, a baked checkerboard or fringe halos in generated cut-outs
  • Recording the source and licence of every shipped asset before release
  • Answering the Steam content survey about AI-generated content from per-asset records

What Makes It Unique

Most image skills stop at the prompt. This one treats a generated image as a draft, requires a hero asset approved at game scale before a family is made, and gates approval on scripted checks plus a look at the game at native resolution. Its gotcha list names concrete failures such as pseudo pixel art, drifting sprite sheets and non-multiple-of-four textures.

Under The Hood

The skill loads a main file and four reference files, plus an art-direction brief and an asset manifest template. It includes two Python scripts that use Pillow: one reads the image files you name and prints a report, the other reads them and writes one contact-sheet image to an output path you choose. Neither script uses the network. The instructions tell the agent to install Pillow with pip when it is missing, to use an image generation or editing tool when one is installed, which may bill per image so the agent should confirm batch size first, and to write manifest and provenance files beside the assets.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers