What does this skill do, and when should you use it?
This skill handles explicit requests for desktop or system screenshots and provides an OS-level fallback when specialized capture tools are unavailable. It supports full-screen, active-window, application, window, pixel-region, and display-oriented capture workflows. Bundled helpers cover macOS, Linux, and Windows, with separate rules for user-specified paths, default screenshot locations, and temporary inspection files. It is a practical fit for repeatable desktop capture, subject to platform permissions and locally available screenshot utilities.
Saves captures to a user-provided path, the operating system's default screenshot location, or a temporary directory; captures macOS applications, windows, window IDs, active windows, regions, or displays; selects scrot, gnome-screenshot, or ImageMagick import on Linux and supports Linux region and active-window capture; runs a PowerShell helper on Windows for full-screen, region, active-window, or window-handle capture; reports the resulting file path.
- A user documenting the complete desktop or multi-display state can take a full-screen capture.
- A user inspecting a desktop application's interface can capture an application or window by name on macOS.
- A user who needs only part of an interface can capture a pixel-defined region.
- A user asking the agent to inspect the current interface can save a temporary screenshot for visual review.
- A user working with a browser or Electron app without integrated capture can use OS-level capture.
How do you install this skill?
- The supplied material omits macos_permissions.swift, macos_window_info.swift, macos_display_info.swift, and the referenced Windows PowerShell helper; confirm they exist after installation.
- Screenshots may expose credentials, personal information, or internal data; confirm the capture scope and output location before use and avoid untrusted destinations.
- The README says the repository is deprecated; confirm that this revision is still the intended maintained source.
- Shell / CLI
- Local filesystem
Python 3PowerShell on Windowsscrot, gnome-screenshot, or ImageMagick import on Linux
The skill is located at skills/.curated/screenshot/. The README says curated skills can be installed in Codex with $skill-installer screenshot, followed by a Codex restart. The source does not state a separate cost. It says an individual skill's license is in its directory's LICENSE.txt, but that file's contents are not included here.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/openai/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/.curated/screenshot" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Ask for a screenshot to use the default save location. For an explicit path, use: python3 <path-to-skill>/scripts/take_screenshot.py --path output/screen.png. For agent inspection, use: python3 <path-to-skill>/scripts/take_screenshot.py --mode temp. For a pixel region, use: python3 <path-to-skill>/scripts/take_screenshot.py --mode temp --region 100,200,800,600. macOS supports --app and --window-name; Linux supports --active-window; Windows uses the PowerShell helper. The installed skill directory path is environment-dependent.
What are this skill's strengths and limitations?
- Covers macOS, Linux, and Windows.
- Supports full-screen, application, window, active-window, window-handle, and pixel-region capture.
- Defines predictable save-location behavior for user requests and agent inspection.
- Includes macOS permission preflight and Linux tool-selection logic.
- macOS application and window capture requires Screen Recording permission.
- Linux requires scrot, gnome-screenshot, or ImageMagick import.
- The supplied material includes no test suite, version requirements, or platform validation record.
- The specific license terms for this skill cannot be confirmed because LICENSE.txt was not provided.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Cross-Platform Screenshot Capture this page ✓ OpenAI · Official | 46 · Use with care | ★ 28k | 4mo ago | — |
| Cua Driver GUI Automation Skill | 61 · Recommended | ★ 30k | 11d ago | MIT |
| Cua Driver Skill | 58 · Recommended | ★ 30k | 11d ago | MIT |
| Cua Spaces (cua-spaces skill) | 46 · Use with care | ★ 30k | 11d ago | MIT |
| Storage Analyzer | 60 · Recommended | ★ 21k | 25d ago | MIT |
The skill explicitly prioritizes specialized capture capabilities such as a Figma MCP/skill or Playwright/agent-browser. It is intended for whole-system captures, explicit desktop-capture requests, or cases where those integrated tools cannot provide the needed image.
How did FollowSkills review this skill?
The scripts save captures to a user-specified path, an OS default location, or a temporary directory, and the macOS preflight explains and requests Screen Recording permission. However, screenshots may contain sensitive screen data, while Linux and Windows lack equivalent confirmation, sensitive-data guidance, scope disclosure, and rollback or cleanup guidance. System commands are also invoked, so points are deducted.
The Python helper includes mutually exclusive argument checks, platform branches, tool detection, and reasonably diagnosable failure messages; its embedded test mode provides limited reproducibility. However, the supplied material does not include the referenced macOS Swift files or the Windows PowerShell helper, and there is no committed comprehensive test suite or CI evidence. The score therefore remains conservative under the static cap.
The skill clearly targets desktop screenshots and documents application/window, region, platform, and tool-priority behavior. Non-fit boundaries, trigger semantics, Chinese-language guidance, and environment prerequisites are incomplete. Its core function is local OS capture and does not depend on overseas online services, so no additional mainland-network deduction is applied.
The documentation is readable and layered, with command examples, platform limitations, multi-display behavior, and error handling. However, the individual skill license is unknown, and versioning, changelog, explicit maintenance ownership, and update path are absent. The README also states that the repository is deprecated, so points are deducted.
The objective is concrete and common modes cover default, temporary, explicit-path, window, and region captures. Yet the supplied evidence omits key referenced helper files, dependency availability is uncertain, and representative outputs were not statically verified. Direct usability is therefore only partially supported.
The source, argument validation, platform-selection logic, and test mode are auditable primary evidence. There are no independent execution records, complete test suite, CI workflow, or third-party corroboration, so the score remains limited.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →