Breaking

Open-Source Skill for Codex Generates 8-Direction Pixel Art From Reference Images

July 1, 2026 at 12:30 EDT

Media Generation
AI Agents
Open Source

An open-source skill called "Character Animation Creator Skill" that automatically generates game-ready 64×64 pixel art characters in eight directions with animation from reference images or text has been released on GitHub. Designed for OpenAI's Codex and GPT Web Agent, it is distributed under the MIT license.

May 8, 2026 · v0.2.0 · MIT License

Text or Image In, Game-Ready Sprite Sheets Out

An open-source "skill" for OpenAI Codex and GPT Web Agent auto-generates 64×64 pixel-art characters across all 8 directions — with idle, walk and attack animations, then cleans and validates every frame.

directions drawn per character

64×64

pixels per animation frame

384×1536

atlas — 6 cols × 24 rows

Frames per animation

Column height = frame count (per direction)

idle

walk

attack

idle is padded to 6 cells · walk & attack are full cycles

Not fully automatic — a 3-step workflow

1 · Generate

AI backend (Codex / GPT imagegen / DALL·E) draws raw frames → generated/

→

2 · Clean up

pixel_snap.py aligns to grid, quantizes palette, sharpens edges → final/

→

3 · Validate

validate_64_sheet.py checks frames, builds contact sheet → qa/

Why it stands out

Built-in post-processing — pixel snapping, palette quantization and validation — cuts the heavy manual burden of hand-drawing 8-direction RPG and top-down animations.

The caveat

Correction scripts compensate for AI's rough edges and inconsistency, but manual adjustment is still sometimes needed — detailed illustration tends to be more stable than pixel art.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.