What happens if no image is found?

I will tell you no image was found in the clipboard and prompt you to copy an image first.

Which OS are supported?

Windows and macOS. The skill runs the Windows PowerShell script or macOS Bash script accordingly.

What next after the image is read?

I will briefly acknowledge what I see in one sentence and wait for your instructions for further analysis.

image-upload

Scanned

npx machina-cli add skill Jobbha/image-uploader-claude-code/image-upload --openclaw

Files (1)

SKILL.md

1.2 KB

Grab the user's clipboard image by detecting their OS and running the correct script:

Windows: powershell -ExecutionPolicy Bypass -File "%USERPROFILE%.claude\clipboard-grab.ps1"

macOS: bash ~/.claude/clipboard-grab.sh

Run the appropriate command for the user's operating system. If you're unsure which OS, check with: uname -s (returns "Darwin" for macOS, or fails/returns something else on Windows).

If the output is a file path (not "NO_IMAGE"), read that image file. Briefly acknowledge what you see in one sentence, then wait for the user's instructions. Do NOT write a detailed description, table, or analysis. The user is sharing the image as context for their next message — treat it like they showed you their screen. Just confirm you see it and ask what they need, or if they included a message with the upload, respond to that.

If the output is "NO_IMAGE", tell the user no image was found in the clipboard and to copy an image first (Win+Shift+S on Windows, Cmd+Ctrl+Shift+4 on Mac, or copy from any app).

$ARGUMENTS

Source

git clone https://github.com/Jobbha/image-uploader-claude-code/blob/main/.claude/skills/image-upload/SKILL.mdView on GitHub

Overview

This skill grabs an image from the user's clipboard and analyzes it. It detects the OS and runs the appropriate script (Windows PowerShell or macOS Bash) to fetch the image, then reads the file path if available. If no image is found, it informs the user and prompts to copy an image first.

How This Skill Works

The skill detects the user's OS and executes the corresponding command: Windows users run a PowerShell script, macOS users run a Bash script. If the command returns a file path (not NO_IMAGE), it reads the image file and briefly acknowledges what you see; if the output is NO_IMAGE, it tells you that no image was found and instructs you to copy an image.

When to Use It

When the user says image upload, paste, screenshot, clipboard, or asks to look at something they copied.
When you need to bring in visual context before discussing details.
When the user wants to share the contents of their screen via clipboard.
When confirming an image is present before performing image-based analysis.
When the assistant should wait for the user's instruction after presenting the image.

Quick Start

Step 1: Copy an image to your clipboard (Windows: Win+Shift+S, macOS: Cmd+Ctrl+Shift+4).
Step 2: Say or request 'image upload' or paste related instruction.
Step 3: I will briefly acknowledge the image and wait for your next instruction.

Best Practices

Test on both Windows and macOS; confirm uname -s identifies the OS correctly.
If a file path is returned (not NO_IMAGE), read and acknowledge the image briefly.
If NO_IMAGE is returned, clearly instruct the user to copy an image first.
Keep the initial response concise to treat the image as context for the next message.
Avoid detailed analysis of the image until the user asks for it.

Example Use Cases

User copies a screenshot of a UI error on Windows and asks what it shows.
User copies a diagram on macOS and wants a quick interpretation.
User asks to look at a screenshot of a network tab from a browser.
User shares a photo from clipboard to confirm what they captured on screen.
User pastes a clip showing a chart and asks for summary.

Frequently Asked Questions

Add this skill to your agents