X Ray logo

X Ray

Community
haorantang97
x-ray

Create and edit X-ray transmission imagery from a brief or visual reference, preserving projection, material attenuation, internal structure, and display polarity across objects, textiles, living subjects, and actions. Use for radiographic visual art and style transfer; assess glass-render, cutaway, collage, and anatomy failures separately.

Overview

Publisherhaorantang97
RepositoryLabArt
Skill namex-ray
Stars
292
Forks
15
Bundled files
6
LicensePolyForm-Noncommercial-1.0.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 6 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by haorantang97 on GitHub. Read the source before you install it.

Installation

Install the X Ray AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/haorantang97/LabArt/tree/main/plugins/x-ray/skills/x-ray
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/haorantang97/LabArt.git /tmp/LabArt
mkdir -p .claude/skills
cp -r /tmp/LabArt/plugins/x-ray/skills/x-ray .claude/skills/x-ray
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable X Ray in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads X Ray on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so X Ray is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

X-ray

Create original images in which external form and internal construction coexist as a transmission projection. Preserve the radiographic mechanism while changing subject, action, or arrangement. Read formation.md before generating; read prompt-recipes.md to construct the prompt.

Define The Image Contract

Record:

  • subject and intended action or relationship;
  • plausible internal structures and materials;
  • one source/detector view;
  • display polarity and monochrome or requested pseudocolor;
  • contrast range, image softness, and framing;
  • reference roles and provenance.

Keep the imaging block separate from the content block. Do not add a fantasy habitat, botanical collage, or architectural cutaway merely because the user requests more creativity. For realistic mode, creativity can come from an unusual subject, interaction, viewpoint, overlap, or captured instant. Use hybrid imagery when requested and identify its mixed mechanism.

Preserve Transmission

  1. Form the image from attenuation along rays through the intact subject. Keep a coherent projection; a front or near-parallel view is an option, not the definition of radiography.
  2. Material composition, energy, density, and path length jointly affect attenuation. Surface color does not determine X-ray darkness.
  3. Overlap accumulates optical depth. Dense or thick regions may hide structure; do not show every component equally clearly.
  4. Choose one display convention throughout the scene. For pale fields, high transmission is light and stronger attenuation darker. The inverted black-field convention is equally usable; record a change of polarity as a changed contract.
  5. Replace optical surface rendering with transmission evidence. Suppress specular glints, glass refraction, studio shading, leather grain, fur photography, floor shadows, and separately lit bones unless the brief specifically requests mixed media.
  6. Make internal topology credible. A detailed skeleton or circuit board is not evidence of correct anatomy or engineering. Obtain appropriate construction/anatomy references when accuracy matters; otherwise report it as unverified.
  7. Treat tint, grain, blur, and fog as display or acquisition choices. Do not use them to conceal failed structure.

Use Visual References

Inspect references before using them. A URL or written description is not visual inspection. Prefer relevant actual radiographs when available, and distinguish them from generated approximations.

Two generated appearance anchors are bundled:

These are synthetic visual examples with known limits, not physics ground truth or validated anatomy. Use an anchor only when its appearance matches the brief. Its subject and layout are not templates for every image. Plugin gallery screenshots are presentation examples, not automatic runtime inputs.

When a tool supports reference images, supply the selected image through its real image-input mechanism and state which properties to preserve. For one-detail changes, start from the same baseline rather than chaining edits that accumulate drift.

Choose The Rendering Route

In Codex, use the system imagegen workflow and built-in image generation. On another platform, use an available raster generator/editor while preserving reference roles and the image contract. Do not install a model, require an API key, or send files to another service without authorization.

If no renderer is available, deliver a renderer-ready prompt and review criteria, and state that no image was generated. If the renderer cannot receive references, state that the reference was inspected but not applied.

Generative imagery is not a physical forward simulation. For quantitative projection fidelity, a CT/volume or material-labelled 3D scene and a suitable forward renderer are needed. Do not claim text-to-image has solved the attenuation equation.

Transfer To Actions And Complex Scenes

Non-still-life requests need a changing pose, contact, pursuit, or interaction. Blurring a static device is an acquisition test, not a dynamic-content test.

  • Trace limbs and joints through overlap; count only what can be justified by the projection.
  • Check every meaningful contact: hand/ball, hand/handlebar, foot/pedal, stick/drum.
  • Check mechanical continuity separately from bodily gesture.
  • Let near/far structures superpose naturally rather than separating them for diagram-like legibility.
  • Use local motion softness only where speed warrants it. Missing fingers, invalid joints, or disconnected mechanisms remain failures.
  • A pose may suggest action without proving its forces or trajectory. Use contact and scene evidence when the action depends on them.

Validate And Repair

Read qa.md. Inspect each output at a useful size and report pass, fail, or uncertain separately for projection, attenuation cues, polarity, visibility, structure, modality, and requested action.

For stability:

  1. Keep the contract and reference fixed.
  2. Change one named content or acquisition condition.
  3. Compare against the baseline.
  4. Repeat relevant cases before claiming reproducibility.

Reference-conditioned consistency, physical correctness, and unrestricted cross-subject stability are different claims. Do not infer any of them from one attractive sample.

If the image becomes a glass render, lit skeleton model, cutaway, or collage in realistic mode, mark that failure even when aesthetically strong. Make a targeted repair, then inspect again. If the repair leaves the issue unresolved, report that limit rather than adding more style adjectives or silently switching modes. Image-conditioned repair can preserve an existing structural error.

Deliver

Show the actual result and save selected images plus exact prompts. Briefly state the intended action, applied reference, and important unresolved structure or modality limits. Keep detailed experimental records separate from runtime resources. Do not market synthetic images as diagnostic scans, physically verified projections, or another creator's work.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the X Ray AI skill do?

Create and edit X-ray transmission imagery from a brief or visual reference, preserving projection, material attenuation, internal structure, and display polarity across objects, textiles, living subjects, and actions. Use for radiographic visual art and style transfer; assess glass-render, cutaway, collage, and anatomy failures separately.

Why use X Ray on TypingMind?

Because you install it once and use it with any model. X Ray is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install X Ray in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/haorantang97/LabArt/tree/main/plugins/x-ray/skills/x-ray. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use X Ray?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with X Ray?

As many as you like. As long as a model supports skills, you can use X Ray with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the X Ray AI skill free?

Yes. It is published on GitHub by haorantang97 under the PolyForm-Noncommercial-1.0.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇