screenshot
$
npx mdskill add SteelMorgan/1c-agent-based-dev-framework/screenshotCaptures desktop screenshots for explicit requests or when tool-specific methods are unavailable.
- Helps users capture full screens, specific apps, windows, or pixel regions on their system.
- Integrates with macOS and Linux via Python scripts and OS-level capture tools.
- Recommends based on user requests, tool availability, and save-location rules.
- Delivers results by saving to user-specified paths, OS defaults, or temporary directories.
SKILL.md
.github/skills/screenshotView on GitHub ↗
---
name: "screenshot"
description: "Use when the user explicitly asks for a desktop or system screenshot (full screen, specific app or window, or a pixel region), or when tool-specific capture capabilities are unavailable and an OS-level capture is needed."
---
# Screenshot Capture
Save-location rules:
1. User specifies path → save there.
2. User asks without path → OS default location.
3. Codex self-inspection → temp directory.
Prefer tool-specific captures (Figma MCP, Playwright, agent-browser) when available. Use this skill for explicit requests, whole-system desktop captures, or when tool-specific capture cannot get what you need.
## macOS permission preflight
Run once before window/app capture. Combine preflight + capture to reduce sandbox prompts:
```bash
bash <path-to-skill>/scripts/ensure_macos_permissions.sh && \
python3 <path-to-skill>/scripts/take_screenshot.py --app "<App>" --mode temp
```
## Python helper (macOS and Linux)
```bash
python3 <path-to-skill>/scripts/take_screenshot.py [OPTIONS]
```
| Option | Example | Note |
|--------|---------|------|
| (none) | | OS default location |
| `--mode temp` | | Codex visual check |
| `--path <path>` | `--path output/screen.png` | Explicit location |
| `--app "<Name>"` | `--app "Codex"` | macOS only, substring match |
| `--app "<Name>" --window-name "<Title>"` | | macOS only |
| `--list-windows --app "<Name>"` | | macOS only, discover window ids |
| `--region x,y,w,h` | `--region 100,200,800,600` | |
| `--active-window` | | Frontmost window |
| `--window-id <id>` | `--window-id 12345` | |
The script prints one path per capture. Multiple windows/displays produce multiple paths with `-w<id>` or `-d<display>` suffixes.
### Linux tool selection (automatic)
Priority: `scrot` → `gnome-screenshot` → ImageMagick `import` → `ffmpeg` (x11grab). For X11 virtual displays, `ffmpeg` is often the most practical option in containers:
```bash
ffmpeg -y -f x11grab -video_size 1920x1080 -i :99 -frames:v 1 /tmp/screen.png
```
Display dimensions: `DISPLAY=:99 xdpyinfo | grep dimensions`
`--app`, `--window-name`, `--list-windows` are macOS-only. On Linux use `--active-window` or `--window-id`.
## PowerShell helper (Windows)
```powershell
powershell -ExecutionPolicy Bypass -File <path-to-skill>/scripts/take_screenshot.ps1 [OPTIONS]
```
| Option | Example |
|--------|---------|
| (none) | Default location |
| `-Mode temp` | Codex visual check |
| `-Path "<path>"` | Explicit |
| `-Region x,y,w,h` | Pixel region |
| `-ActiveWindow` | Ask user to focus first |
| `-WindowHandle <id>` | Specific window |
## Direct OS fallbacks
### macOS
```bash
screencapture -x output/screen.png # full screen
screencapture -x -R100,200,800,600 output/region.png # region
screencapture -x -l12345 output/window.png # window id
```
### Linux
```bash
scrot output/screen.png # full
scrot -a 100,200,800,600 output/region.png # region
scrot -u output/window.png # active window
import -window root output/screen.png # ImageMagick full
ffmpeg -y -f x11grab -video_size 800x600 -i :99+100,200 -frames:v 1 output/region.png
```
## Error handling
- macOS sandbox errors ("screen capture blocked", `ModuleCache`) → rerun with escalated permissions
- macOS no matches → `--list-windows --app "Name"` → retry with `--window-id`
- Linux tool missing → `command -v scrot`, `command -v import`
- Always report saved file path