Publisher: OpenAI · Editorial profile by AI Tools Ratings · Review method
Not runtime tested; source-document review only. No security certification.
What task is this skill designed for?
Collect a screenshot of a local application defect so a teammate can see the exact dialog, clipping or display state being discussed.
What input and output should you expect?
Input: The application or window to capture, the relevant screen region and an explicit destination for the image. Close or obscure unrelated private windows before capture.
Output: Expected, not independently tested: an image saved to the requested location, with the selected application state visible. This is evidence capture rather than an explanation of the defect.
What does it need to run?
An agent with operating-system command access and a desktop session. The source includes Python and PowerShell paths and platform-specific capture dependencies. On macOS, Screen Recording permission may be needed; a headless server cannot provide an ordinary desktop screenshot.
Authored for Codex in the OpenAI skills repository. This directory has not independently installed or executed the skill.
How should you approach installation?
- Inspect the pinned source and the helper appropriate to your operating system.
- Install the skill through your agent’s documented local-skill mechanism; check its desktop and capture permissions first.
- Use a non-sensitive test window and verify the saved image before trying a real support case.
Installation guide: discovery, dependencies and runtime checks
What files, accounts or data can it access?
Screen capture can expose information from unrelated windows, browser tabs and notifications. Inspect the saved image before sharing it and use the smallest region that proves the issue. Permission to capture a screen is not permission to upload the image.
What does the skill cost?
Source instructions are available without a listed skill purchase; the agent environment and any required platform software have their own terms.
What are its limits and alternatives?
A screenshot records a moment, not the full interaction sequence. It cannot prove keyboard accessibility or that a workflow completed. Multi-monitor layouts, display scaling and missing system permissions can change the result.
Use a browser or design-tool screenshot API for content already accessible through that tool. Choose an OS-level capture only when the relevant interface is outside those surfaces.
What could you ask in a small trial?
Try this with a small sample:
Capture only the test application window showing the clipped settings dialog. Save the image locally, tell me the path, and do not upload or include unrelated windows.
How would you verify the result?
- Open the image and confirm the intended window and dialog are visible.
- Check that text is readable at the saved resolution and no confidential background is included.
- Record the reproduction steps separately; do not treat the image as proof of a completed fix.
Use a baseline and record acceptance evidence before treating a successful task as a reliable workflow.
Which source and license were checked?
Read the reviewed SKILL.md at its pinned commit · Complete reviewed folder
Repository and current publisher guidance
Checked:
Apache-2.0 — identified in skills/.curated/screenshot/LICENSE.txt at the reviewed commit. Third-party assets and dependencies may have separate terms. License evidence
Version details and review scope
Source: openai/skills/skills/.curated/screenshot
Reviewed commit: 49f948faa9258a0c61caceaf225e179651397431
SKILL.md SHA-256: 081935a6a163277537d46365f49d6b4a3cb40b4748347e7e88759c5927fa8cf5
Source-document review. Source text and stated dependencies were inspected for this profile. Referenced code was not comprehensively audited, no runtime test was performed, and source files are not mirrored here.