Skip to content

nddipiazza/robos / .agents/skills/jrpg-generate-art

jrpg-generate-art

Verified

Generate background maps, dungeons, battle backdrops, character sprites and menu screens for a parsed JRPG kgraph with a Gemini image model, using the collected resources as references.

Generate JRPG Art (jrpg-generate-art)

Generates images for a game already parsed with jrpg-parse-assets and adds them to its kgraph.

Input

$ARGUMENTS — the parsed output folder, and optionally which kinds of image to generate.

Requirements

  • agy signed in. Images use its generate_image tool through packages/robos-lib/ai-image.js, so no API key is needed.
  • python3 with Pillow, for fitting images to the game's size and colors.

Procedure

  1. See what would be generated. text-only jobs have no reference image, so the model invents them from the guide's text:
    node packages/robos-jrpg-parser/bin/jrpg-parser.js plan --out <dir>
    
  2. Try a few first, and look at the results before running everything. Each image takes about a minute:
    node packages/robos-jrpg-parser/bin/jrpg-parser.js art --out <dir> --limit 3
    
  3. Run the rest. Finished jobs are skipped, so this also resumes an interrupted run:
    node packages/robos-jrpg-parser/bin/jrpg-parser.js art --out <dir> [--kinds world-map,area-map] [--force]
    

If one image is wrong, look at its raw file first. Redo just that image with --only <id>. If the raw is fine and the shrink step spoiled it, change tools/nesify.py and run art --out <dir> --refit, which does not call the model.

Kinds: world-map, area-map, battle-backdrop, class-sprite, menu-screen, and enemy-sprite, which is off by default because a game has over a hundred enemies.

Output

  • assets/generated/<kind>/<id>.png — the image fitted to the game. Maps keep the model's size and colors with noise removed, sprites and backdrops are game size on the NES palette, and menus are game size in their own colors
  • assets/generated/raw/<id>.img — what the model returned
  • assets/generated/manifest.json — model, prompt hash, and references for each image
  • A jrpg:GeneratedImage node for each image, with jrpg:depicts, jrpg:references and jrpg:basis (reference or text-only)

Rules

  • Generating sends the reference images to Google through agy. Only run it on files the person gave you for this purpose.
  • This art derives from a commercial game. Keep the output private unless the person says to publish it, and then publish only with jrpg-parser publish.
  • Never publish or commit what the person downloaded: source files, working folders, or anything quoted from them.
  • Do not use --dangerously-skip-permissions. Headless agy can read files and generate images without it.

Validation

  • Open a few images from each kind and check them by eye.
  • jrpg-parser coverage --out <dir> still runs, and failed jobs are listed at the end of the run.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 197kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall