Does it infer functionality from UI elements?

The skill follows strict fidelity rules to describe only what is visible. For example, it will describe a 'button labeled Save' rather than assuming 'this button saves the file'.

How does the skill handle SVG files differently than bitmaps?

The skill uses a dual-mode approach for SVGs: it views the file visually to understand layout and simultaneously reads the SVG source code to extract precise labels, IDs, and embedded text.

What happens if a screenshot contains truncated code?

The model extracts only the visible code verbatim and explicitly notes in the summary that the content appears truncated at the top, bottom, or edges.

Can this skill identify specific people in photographs?

No. To maintain high fidelity and privacy, the skill is prohibited from identifying individuals by name or speculating on emotions unless explicit text labels are present in the image.

Image Summarization

Name: Image Summarization
Author: Jamie-BitFlight

byJamie-BitFlight

•

学习与文档

Analyzes and documents visual content including UI screenshots, architecture diagrams, and terminal outputs using multimodal tools.

The Image Summarization skill provides a specialized methodology for Claude Code to interpret and document visual media with high fidelity. It supports a wide range of formats including PNG, JPG, and SVG, applying distinct strategies for UI layouts, cloud architecture diagrams, data charts, and code snippets. By prioritizing visible evidence over inference, this skill ensures that technical documentation, bug reports, and system overviews generated from images are accurate, structured, and include essential metadata like confidence scores and source paths.

主要功能

01Structured YAML metadata output with confidence assessments

02Specialized analysis strategies for UI screenshots, diagrams, and charts

0320 GitHub stars

04Multimodal support for PNG, JPG, WebP, and SVG formats

05Dual-mode SVG processing (visual rendering + source text extraction)

06Verbatim extraction of code and terminal output from screenshots

使用场景

01Generating technical descriptions and accessibility labels for UI components

02Analyzing terminal output and error screenshots for troubleshooting

03Documenting legacy system architectures from inherited diagram images

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jamie-bitflight/claude_skills image-summarization

For use in Claude.ai and ChatGPT

Download Skill