screenshot

Captures screen images, windows, and clipboard content for visual debugging and UI verification.

Updated Aug 2, 2026
One-click install
npx skills add https://github.com/jeje-it/template-projet --skill screenshot-jeje-it
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: screenshot
Source: https://github.com/jeje-it/template-projet/tree/main/.agents/skills/screenshot
Command: npx skills add https://github.com/jeje-it/template-projet --skill screenshot-jeje-it

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When debugging visual issues or verifying UI changes, an AI assistant cannot see what you see on your screen. This Skill lets the assistant capture your screen, specific windows, or clipboard images so it can answer questions about visible content and diagnose layout problems. ## Core Features & Use Cases - Fullscreen and Window Capture: Take screenshots of the entire screen or target a specific window by name, with automatic focus restoration. - Resolution Control: Choose between quarter, half, high, and full resolution to balance token usage against detail level. - Clipboard Image Reading: Read images the user has copied to the clipboard for immediate analysis. - Use Case: You just changed a CSS layout and something looks off. Ask the assistant to capture your screen, and it will inspect the rendered page, identify the misaligned element, and relate it back to your code. ## Quick Start Ask the assistant to take a screenshot of your screen and describe what is wrong with the current layout.

Frequently Asked Questions about screenshot

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

FAQPage Schema
How do I let an AI assistant see my screen for debugging?▼

Use the screen capture tool to take a screenshot, which the assistant then analyzes to answer questions about visible content. The default mode captures the fullscreen at half resolution, balancing detail against token usage.

How do I capture a specific window instead of the full screen?▼

First list all open windows, then call the capture tool with a window name substring. The target window is activated and restored if minimized, captured, and then focus returns to the previous window.

What screenshot resolution should I use to save tokens?▼

Use half resolution for general layout checks at roughly 691 tokens for 1080p, or quarter resolution at about 173 tokens when only checking presence. Reserve high or full resolution for reading small text or pixel-perfect comparisons.

Can the assistant read an image I copied to the clipboard?▼

Yes, the clipboard tool with the readImage action reads images currently stored on the clipboard. This is useful when you have already copied a screenshot or image and want the assistant to analyze it.

Why does my screenshot use too many tokens?▼

Full resolution captures of a 1080p screen consume around 2,765 tokens. Switch to half or quarter resolution for layout overviews, and only use full resolution when fine detail is required.