Menu
Use this to capture screen context.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Use when the user explicitly asks for a CI check or to push their branch — e.g. "ci readiness", "check ci", "pre-push check", "ready for CI", "ci check", "ready to push", "push my changes", "push the branch", "let's push". Catches common CI failures before pushing — formatting, stale API reports, missing changesets, policy violations.
Fluid Framework client release group — minor releases, patch releases, and post-release type test updates. Covers release prep, branching, version bumps, changelogs, release notes, and type test baselines. In autonomous mode, auto-detects state from the schedule and repo, attempts to execute, and falls back to a GitHub issue on failure. Triggers on "release", "do the release", "release status", version bump, release notes, changelog, release branch, or release engineering.
IMMEDIATELY USE THIS SKILL when creating or develop anything and before writing code or implementation plans - refines rough ideas into fully-formed designs through structured Socratic questioning, alternative exploration, and incremental validation
Use when implementing user interfaces or user experiences - guides through exploration of design variations, frontend setup, iteration, and proper integration
Use this skill when faced with a difficult debugging task where you need to replicate some bug or behavior in order to see what is going wrong.
Use when you need to create a new custom skill for a profile - guides through gathering requirements, creating directory structure, writing SKILL.md, and optionally adding bundled scripts
| name | using-screenshots |
| description | Use this to capture screen context. |
You CAN take screenshots by combining the Bash tool with platform-specific screenshot commands. Screenshots are saved as image files, then loaded into your context using the Read tool for visual analysis.
Use this skill when I ask you to:
| Platform | Command | Interactive Selection |
|---|---|---|
| macOS | screencapture | -i flag (area selection) |
| Linux | gnome-screenshot, scrot, or import | -a or -s flag |
Standard workflow:
uname -s/tmp/screenshot_$(date +%s).pnguname -s
Darwin for macOSLinux for LinuxmacOS: screencapture is always available (built-in)
Linux: Check in priority order:
which gnome-screenshot || which scrot || which import || echo "none"
Priority order (best compatibility):
gnome-screenshot - works on both X11 and Waylandscrot - lightweight, X11 onlyimport - part of ImageMagickUse timestamped filename to avoid conflicts:
macOS:
screencapture -i /tmp/screenshot_$(date +%s).png
-i enables interactive area selectionLinux with gnome-screenshot:
gnome-screenshot -af /tmp/screenshot_$(date +%s).png
-a for area selection-f specifies filenameLinux with scrot:
scrot -s /tmp/screenshot_$(date +%s).png
-s enables selection modeLinux with import:
import /tmp/screenshot_$(date +%s).png
Read tool: file_path="/tmp/screenshot_12345.png"
The Read tool displays images visually. You'll see the screenshot and can analyze it.
Once loaded, you can:
rm /tmp/screenshot_12345.png
Only remove if I won't need the file again.
If no screenshot tool is available on Linux:
Inform me which tool is missing
Suggest installation:
sudo apt install gnome-screenshotsudo dnf install gnome-screenshotsudo pacman -S gnome-screenshotsudo apt install imagemagick (or equivalent)Alternative: Ask user to manually take screenshot and provide path
Reality: You CAN via Bash + screenshot CLI tools
Reality: The screenshot file must be loaded with Read tool to see it
Reality: Always use absolute paths (/tmp/...) for Read tool
Reality: Must detect which tool is installed before attempting capture
User: "Take a screenshot and help me debug this UI bug"
1. Check platform:
uname -s → Linux
2. Check available tools:
which gnome-screenshot → /usr/bin/gnome-screenshot
3. Capture screenshot:
gnome-screenshot -af /tmp/screenshot_1729012345.png
→ User selects area, file saved
4. Load into context:
Read: file_path="/tmp/screenshot_1729012345.png"
→ Image displays visually
5. Analyze:
"I can see the button alignment is off. The 'Submit' button
is 5px lower than the 'Cancel' button..."
6. Optional cleanup:
rm /tmp/screenshot_1729012345.png