hatch-pet

Create and package Codex-compatible v2 animated pet spritesheets from art or brand references.

1|Updated Aug 11, 2026
One-click install
npx skills add https://github.com/Chia1104/agent-air --skill hatch-pet-chia1104
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hatch-pet
Source: https://github.com/Chia1104/agent-air/tree/main/skills/codex/hatch-pet
Command: npx skills add https://github.com/Chia1104/agent-air --skill hatch-pet-chia1104

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Pillow, and includes scripts (resource) and references (resource) components.

What problem does it solve? Creating a custom animated pet for Codex requires producing a precise 8x11 sprite atlas with 9 animation rows and 16 look directions, plus deterministic validation and packaging—an error-prone process to do manually. ## Core Features & Use Cases - End-to-End Pet Generation: Generate a base character, 9 standard animation rows (idle, running, waving, jumping, failed, waiting, review), and 16 clockwise look directions, then assemble them into a 1536x2288 atlas packaged with spriteVersionNumber 2. - Repair and Upgrade Workflows: Validate existing 8x9 or 8x11 atlases, repair failed rows or look cells, and upgrade older pets to the v2 format. - Brand-Inspired Mascots: Run a lightweight web-research discovery step to turn a company or product name into visual cues for a brand-inspired pet. - Use Case: Ask for a plush-style mascot based on your startup's brand, and the Skill researches the brand, generates the character, produces all animation rows, runs QA, and installs a ready-to-use pet package under ~/.codex/pets/. ## Quick Start Ask the agent to hatch a new Codex pet named "Pixel" in a clay style from a short description or reference image.

Frequently Asked Questions about hatch-pet

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a custom animated pet for Codex?▼

Provide a name, description, style, or reference images and the Skill generates a base character, nine animation rows, and sixteen look directions. It then assembles a 1536x2288 atlas and packages it with spriteVersionNumber 2 under ~/.codex/pets/.

What image format does a Codex v2 pet spritesheet require?▼

A v2 pet uses an 8-column by 11-row atlas of 192x208 cells, totaling 1536x2288 pixels, saved as PNG or lossless WebP with a transparent background. The pet.json manifest must set spriteVersionNumber to 2 or the spritesheet is rejected.

Can I make a pet from my company or product brand?▼

Yes. When you give only a brand or company name, a discovery worker researches official sources for palette, tone, and motifs, then produces a compact brief that guides pet generation without copying logos or text.

Can hatch-pet repair an existing pet spritesheet?▼

Yes. It validates existing 8x9 or 8x11 atlases, preserves approved rows, and regenerates only failed rows or look cells before deterministic reassembly. A valid 8x9 atlas can be upgraded by generating rows 9-10 and packaging as v2.

Why does the Skill require the bundled Python runtime?▼

The deterministic scripts depend on Pillow for atlas assembly, chroma-key removal, frame extraction, and despill cleanup. You must call load_workspace_dependencies and use the returned Python path; a bare system python is not supported.

Does hatch-pet generate images directly through an image API?▼

No. All visual generation is delegated to the installed imagegen skill, which chooses its own built-in-first path and fallback rules. The bundled scripts handle only deterministic image processing such as extraction, validation, and atlas composition.