PocketMen with You skill
A Codex skill that builds an animated companion from two or more of your photos and generates it on your own machine
- 295 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Processes personal photos of people and animals, and the finished companion mirrors their appearance
- Runs local Python scripts and downloads model weights from outside hosts
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/pocketmen-with-youDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Processes personal photos of people and animals, and the finished companion mirrors their appearance
- Runs local Python scripts and downloads model weights from outside hosts
This is third-party code. Review the repository files before installing.
What it does
PocketMen with You builds a companion pet for Codex from two or more of your photos. First the agent studies the pictures and writes down stable traits: head shape, coat, eye color, accessories. The skill then generates a master look and key poses for nine states, from idle and running to jumping and failed. The result is an 8 by 9 sprite atlas that the skill installs into Codex right away. Generation runs on your machine. The main engine uses the open flux2-klein-4b model under Apache-2.0, and for the closest likeness you can add Qwen-Image-Edit-2511. Without a suitable GPU the skill falls back to a deterministic engine. It only moves the source pixels and invents no new poses. There are three quality profiles: draft, balanced and max. There are five styles, from close-to-photo to plush and an original capsule creature. From a terminal, `pocketmen create` builds a companion.
Who it is for. For Codex users who want a companion that looks like their pet, a friend or themselves.
Good fit when
- You want a Codex companion shaped like your cat, dog or yourself from two or three photos
- An OpenAI key is unavailable or you do not want to send personal pictures to an outside service
- You have an NVIDIA GPU with plenty of video memory and want neural pose generation over simple pixel animation
- You want stylization: toy-like chibi, plush or an original capsule creature
Not a fit when
- You have fewer than two photos: the skill needs at least two
- You want ordinary illustrations or general image generation: the engine targets the Codex companion atlas
- You lack rights to the images of people or characters you want to build from
- You expect proprietary cloud editor quality in any scene: open models fall behind them on complex scenes
Example request
Use $pocketmen-with-you to create a companion from these two photos of my dog, style soft-real, quality balancedLimitations
The neural engine targets an NVIDIA GPU with about 13 GB of video memory, and other machines get a simpler deterministic engine. Model weights download from their original hosts into the Hugging Face cache and are not in the repository. The models carry their own licenses, and the skill avoids the FLUX.2 dev variant because of its non-commercial license. Open models fall behind proprietary ones on general knowledge, text in images and complex scenes. A ready INSTALL_SKILL.bat installer exists only for Windows.
How to disable. Delete the pocketmen-with-you skill folder from your skills directory and the generated companion from the Codex pets folder. Downloaded weights can be cleared from the Hugging Face cache.
FAQ
Do I need an OpenAI key for PocketMen with You?
No. The skill works without OPENAI_API_KEY and does not call the hatch-pet helper skill.
What GPU does the neural engine need?
The skill targets NVIDIA with about 13 GB of video memory. pocketmen doctor runs the check. The Windows installer detects memory itself and installs the matching dependency set.
What comes out?
An 8 by 9 atlas of 192 by 208 pixel cells, 1536 by 1872 overall. The installed package holds only pet.json and spritesheet.webp, and the source photos stay with you.
How do draft, balanced and max differ?
Draft makes one neural master look plus deterministic animation. Balanced adds separate key poses per state. Max generates a key pose for each of the nine states independently.
Can I build a companion from someone else's photo?
The skill is meant for your own pictures. You need rights to images of people and characters, and personal photos should stay out of Git, releases and issues.
Discussion
No questions or reviews yet.
Related
SkillsA set of skills for Claude Code and Codex that assemble detailed image prompts from an element library and composition schemes
A Codex skill that redraws your photo as a watercolor illustration on white paper, with wide margins and a handwritten caption
A skill that turns your photo or description into a hand-drawn character and uses it for diagrams, knowledge cards and slides
A skill for Claude Code and Codex that builds cinematic product videos with Remotion from shot recipe cards, a ready template and audio assets
A skill that writes prompts for video, image and music generation on 14 platforms and can submit them in the browser
Sprite Gen
sprite-gen
A skill for Codex and Claude and a Python CLI: builds 2D game sprites from one base image, removes the background and assembles animation atlases
Comparison with similar tools
Entries that share the main search query, and how they differ.
README in short
The README shows the pipeline from photos to a Codex install and compares the two models. It ends with privacy and license sections and a note that the project is not affiliated with Nintendo, OpenAI or the other companies it mentions.
README badge
Original SKILL.md
--- name: pocketmen-with-you description: Turn 2+ user reference images into a high-fidelity animated Codex companion using PocketMen's own local stack. Prefer the open-weight Neural Local Studio (FLUX.2 klein 4B; optional Qwen-Image-Edit-2511 Identity-Max) on compatible hardware; otherwise fall back to the deterministic local renderer. Never require hatch-pet or OPENAI_API_KEY for the normal workflow. --- # PocketMen with You — Neural Local Studio Create a Codex companion from **at least two** user-provided reference images. The normal workflow is self-contained and **must not invoke `$hatch-pet` or request `OPENAI_API_KEY`**. ## Product goal The user should be able to upload 2+ images and get a polished, identity-consistent Codex companion with minimal setup. PocketMen should aim for hosted-image-editor-like quality **within this narrow companion-creation task**, while being honest that local open-weight models are not guaranteed to equal a proprietary frontier model on every visual task. ## Engine order Use this decision order: 1. inspect the references visually and build an Identity Lock; 2. run PocketMen hardware doctor; 3. if compatible neural runtime/hardware is available, use `neural-local`; 4. if the neural runtime is missing but an NVIDIA GPU with roughly 13 GB+ VRAM is present, bootstrap the skill-local neural profile once, then retry; 5. if neural inference is unavailable or fails, fall back to the deterministic local renderer; 6. never route failure into hatch-pet or OpenAI API-key setup. ### Default neural backend `flux2-klein-4b` - open-weight local generation/editing; - multi-reference editing; - commercial-friendly Apache-2.0 weights; - intended for consumer NVIDIA GPUs; - preferred for most users. ### Identity-Max backend `qwen-image-edit-2511` Use only when the user prioritizes identity consistency over setup size/latency or when FLUX state outputs visibly drift. This backend is heavier and should not be silently installed unless needed.