Visual To Spec

Community

Jason904

visual-to-spec

Use when a user provides a reference UI screenshot and asks to extract visual specs, generate design tokens, create a component tree, or produce a frontend implementation contract, including dashboards, landing pages, mobile screens, image-led UIs, panorama scenes, 3D exhibit UIs, and screenshot-to-spec tasks.

Overview

Publisher	Jason904
Repository	`ui-skill-lab`
Skill name	`visual-to-spec`
Stars	144
Forks	9
Bundled files	10
License	MIT
Links	Skill source Repository

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.
10 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by Jason904 on GitHub. Read the source before you install it.

Installation

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

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.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/Jason904/ui-skill-lab/tree/main/.codex/skills/visual-to-spec
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.

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/Jason904/ui-skill-lab.git /tmp/ui-skill-lab
mkdir -p .claude/skills
cp -r /tmp/ui-skill-lab/.codex/skills/visual-to-spec .claude/skills/visual-to-spec

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Visual To Spec 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 Visual To Spec 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 Visual To Spec 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.

Version: 2.0.0

visual-to-spec

Convert a finalized UI screenshot into a complete frontend implementation contract.

Trigger Conditions

User provides a UI screenshot and asks to extract visual specs
User asks to generate design tokens from a screenshot
User asks to build a component tree or frontend implementation contract from a screenshot
User asks to analyze screenshot layout, color, typography, spacing, overlay controls, or panorama/3D scene UI
Keywords: "extract spec from screenshot", "image to design spec", "screenshot to spec"

Required Inputs

Item	Description
reference.png	Finalized UI screenshot (the only required input)
Project token standards	Optional, adapt to existing standards if available

Outputs

All files written to 03_visual_spec/ directory:

File	Content
visual-analysis.md	UI type, visible element evidence, composition, color, typography, spacing analysis
layout-spec.md	Layout positioning and sizing spec, branching by UI type
component-tree.md	Hierarchical component tree with per-component style summaries
tokens.json	DTCG-inspired design tokens with source/confidence annotations
implementation-risks.md	Implementation risk assessment
human-review-needed.md	Low-confidence items and manual review checklist
DESIGN.md	Summary document referencing all above files

See references/output-files.md for detailed output formats.

Source of Truth

reference.png is the sole visual source of truth.

Do not redesign.
Do not treat UI elements invisible in the screenshot as screenshot facts.
Do not introduce colors, fonts, spacing, panels, charts, or navigation structures absent from the image.
Do not default to a dashboard / SaaS / three-column layout; UI type determination must come first.
reference-visible components must have bbox and visual evidence in the screenshot; components without bbox must not be labeled reference-visible.
Inferred implementation wrapper components (e.g., AppShell, MainContent, SceneImageLayer, OverlayLayer) are permitted but must be labeled source: inferred-implementation.
For image-led, panorama-scene, 3D exhibit, and product-render screenshots, output an asset strategy: which visual content must be raster/image assets, and which can be implemented as HTML/CSS/SVG.

All values must carry source and confidence annotations. See references/source-strictness-rules.md for complete rules.

Phase Overview (9 phases, sequential)

#	Phase	Key Output
1	Global composition analysis + UI type determination + visible element evidence	visual-analysis.md §1
2	Color extraction	visual-analysis.md §2 + tokens.json (color)
3	Typography extraction	visual-analysis.md §3 + tokens.json (typography)
4	Spacing and grid system	visual-analysis.md §4 + tokens.json (spacing)
5	Layout specification	layout-spec.md
6	Component tree construction	component-tree.md
7	Token merge and validation	tokens.json (final)
8	Implementation risk assessment	implementation-risks.md
9	Human review checklist	human-review-needed.md

See references/extraction-workflow.md for detailed workflow.

Hard Rules

Every extracted value must include source, confidence, confidenceLabel, and strictness.
reference-visible must include bbox or reference bbox within the same component detail; visibility claims without bbox are invalid.
Phase 1 must output uiType, one of: application-dashboard, panorama-scene, image-led-landing, mobile-app, poster-like-ui, unknown.
If uiType is panorama-scene or image-led-landing, phases 5/6 must use the image layer + overlay layer layout model.
Components not present in the screenshot must be listed in Rejected Assumptions, not placed in the component tree.
Values with confidence < 0.6 must appear in human-review-needed.md; low-confidence layout type must go into human review even if confidence >= 0.6.
Every color in the component tree must exist in tokens.json.
Every component dimension must match layout-spec.md.
Sum of sibling component widths must not exceed parent container width; overlay components must verify bbox does not overflow.
Typography scale must decrease monotonically: h1 > h2 > h3 > body.
No orphan tokens (defined but unused).
No missing tokens (used in component tree but undefined).

Capability Boundaries

See references/script-backed-boundary.md for detailed boundary definition.

Category	Examples
LLM can do	Visual recognition, UI type determination, color estimation, layout judgment, component classification, proportion estimation
Script-assisted	Precise color values, WCAG contrast, delta-E, exact pixel measurement, bbox measurement, visual regression comparison
Human confirmation	Low-confidence colors, ambiguous component boundaries, uncertain font identification, whether the main visual needs real assets

When scripts are unavailable, LLM does best-effort estimation and annotates confidence and notes.

Supporting Files Index

references/

extraction-workflow.md — 9-phase detailed workflow, UI type branching, asset strategy rules
token-schema.md — Token JSON schema definition and field descriptions
output-files.md — Structure templates for each output file
source-strictness-rules.md — Complete rules for source/confidence/strictness/bbox labels
script-backed-boundary.md — Boundary definitions for LLM / script / human operations

examples/

tokens.example.json — Complete tokens.json example
component-tree.example.md — Component tree example, including panorama-scene vs dashboard branching
layout-spec.example.md — Layout spec example, including overlay coordinate model
visual-analysis.example.md — Visual analysis example, including UI type, visible elements, rejected assumptions
human-review-needed.example.md — Human review checklist example

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 Visual To Spec AI skill do?

Why use Visual To Spec on TypingMind?

Because you install it once and use it with any model. Visual To Spec 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 Visual To Spec in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Jason904/ui-skill-lab/tree/main/.codex/skills/visual-to-spec. 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 Visual To Spec?

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 Visual To Spec?

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

Is the Visual To Spec AI skill free?

Yes. It is published on GitHub by Jason904 under the MIT 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

Grill Me

mattpocock

CommunityPopular

A relentless interview to sharpen a plan or design.

Brand Discovery

Use when a brand needs to discover or articulate its identity through structured multi-session interviews. Covers purpose, positioning, audience, personality, voice, narrative, and founder-brand…

Competitive Platform Analysis

affaan-m

CommunityPopular

Use when scoping a competitive landscape — identifying, categorising, and score-filtering a competitor set before any benchmarking begins. Decides who counts as a competitor, which tier they belong…

Frontend Slides

Create stunning, animation-rich HTML presentations from scratch or by converting PowerPoint files. Use when the user wants to build a presentation, convert a PPT/PPTX to web, or create slides for a…

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

visual-to-spec

Trigger Conditions

Required Inputs

Outputs

Source of Truth

Phase Overview (9 phases, sequential)

Hard Rules

Capability Boundaries

Supporting Files Index

references/

examples/

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now