skills.rest
giszzt
giszztCOMMUNITY
Explore all skills by this author →
GitHub

figecho

Extract visual style systems from reference diagrams and generate style-consistent technical figures.

34|3|Updated Aug 25, 2026
Search Tags:#image generation#architecture diagram#style prompt#style library#diagram style transfer#technical figure#visual fingerprint
One-click install
npx skills add https://github.com/giszzt/FigEcho --skill figecho-giszzt
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: figecho
Source: https://github.com/giszzt/FigEcho
Command: npx skills add https://github.com/giszzt/FigEcho --skill figecho-giszzt

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve? Recreating the visual style of a technical diagram for new content is hard: style is more than colors, and ad-hoc prompts produce generic results. FigEcho turns a reference diagram into a reusable, evidence-backed visual fingerprint and a complete modular prompt, so new architecture diagrams, roadmaps, and method figures come out in the same visual family. ## Core Features & Use Cases - Visual Fingerprint Extraction: Distills 5-7 high-distinctiveness style anchors across layout, components, connectors, palette, typography, and iconography, each with evidence and transfer rules. - Three Delivery Modes: prompt-only (standalone style prompt), analysis-and-prompt (style report plus prompt), and image-and-prompt (final PNG plus the exact generation prompt sent to the model). - Style Library: Save, search, rename, validate, and reuse named Style Packs that bundle the full prompt, style spec, and original reference images; ships with 6 example packs. - Use Case: Give it a paper's method-overview figure and your own module list, and it plans the new diagram's topology, fills the style prompt with your content, binds the reference image to an image backend (built-in tool, Labnana, OpenAI, or Gemini), and verifies the render against each style anchor. ## Quick Start Extract the layout, colors, typography, and connector style from this reference architecture diagram and give me a reusable style prompt.

Frequently Asked Questions about figecho

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

FAQPage Schema
How do I copy the style of a technical diagram for new content?▼

Provide the reference diagram and your new content. The skill extracts a visual fingerprint of 5-7 style anchors, compiles a modular style prompt, plans the new diagram's layout, and generates a style-consistent PNG with the reference image bound to the generation call.

What image generation backends does it support?▼

It prefers the agent's built-in reference-capable image tool. Without one, scripts/generate_figure.py tries Labnana GPT-Image-2, Labnana Nano Banana, OpenAI official, Gemini official, then a custom configured command, using keys from a .env file.

Can I save and reuse a diagram style later?▼

Yes. Styles are saved as named Style Packs containing the full prompt, style spec, and original reference images. You can search the local Style Library by name, alias, or use case, and packs graduate from candidate to validated after passing rendered and cross-topic portability checks.

Does it work without a reference image?▼

It first searches the Style Library and asks you to confirm a matching template. Only with your explicit approval does it create an original style, and the delivery clearly states the style was designed without reference evidence.

What are the limitations of style transfer for diagrams?▼

It transfers layout grammar, component rules, color relationships, typography categories, and icon abstraction, but never source text, logos, data, watermarks, or proprietary content. It is not for pixel-level copying, watermark removal, or simple diagrams with no style intent.

Related Skills

Explore More Design & Creative Skills→
spencerbeggsspencerbeggs

design-link

Build cross-reference graphs from design documents and output Mermaid, text, or JSON.

Community
Advanced
wanshuiyinwanshuiyin
15.6k

figure-spec

Generate deterministic publication-quality architecture and workflow diagrams from JSON specs into editable SVG.

Community
Intermediate
mahg-esmahg-es
3

visual-identity

Design and maintain a visual identity system with logo, color, typography, iconography, and imagery guidelines.

Official
Intermediate
sumik5sumik5
1

creating-diagrams

Create flowcharts, sequence diagrams, and ER diagrams using Mermaid syntax.

Community
Advanced