Agent Skills › n0an/VivaDicta › screenshot

screenshot

GitHub

通过 xcrun simctl 命令捕获 iOS 模拟器截图，支持保存路径报告及基于指令的视觉分析、元素检查或文本提取。

.agents/skills/screenshot/SKILL.md n0an/VivaDicta

Trigger Scenarios

需要获取 iOS 模拟器当前屏幕画面需要验证 UI 界面显示状态需要分析截屏内容中的特定元素或文本

Install

npx skills add n0an/VivaDicta --skill screenshot -g -y

More Options

Non-standard path

npx skills add https://github.com/n0an/VivaDicta/tree/main/.agents/skills/screenshot -g -y

Use without installing

npx skills use n0an/VivaDicta@screenshot

指定 Agent (Claude Code)

npx skills add n0an/VivaDicta --skill screenshot -a claude-code -g -y

安装 repo 全部 skill

npx skills add n0an/VivaDicta --all -g -y

预览 repo 内 skill

npx skills add n0an/VivaDicta --list

SKILL.md

Frontmatter

{
    "name": "screenshot",
    "description": "Capture screenshots from iOS Simulator using native xcrun simctl command for documentation, debugging, and visual verification. Optionally analyze the screenshot.",
    "disable-model-invocation": true
}

screenshot

You are given the following context: $ARGUMENTS

Task

Capture a screenshot of the iOS Simulator using xcrun simctl and optionally analyze it based on user instructions.

Instructions

Create screenshots directory if needed:
```
mkdir -p llmtemp/screenshots
```

Capture the screenshot:

TIMESTAMP=$(date +%Y%m%d_%H%M%S)
xcrun simctl io booted screenshot llmtemp/screenshots/screenshot_${TIMESTAMP}.png

Handle the optional argument:
- If $ARGUMENTS is empty or not provided: Capture and save, then report the saved path
- If $ARGUMENTS is provided: After saving, read the screenshot and follow the requested instruction
Common argument patterns:
- "take a look" / "analyze" / "what's on screen" - Read and describe what's visible
- "check for [element]" - Look for specific UI elements
- "is there [something]" - Verify presence of specific content
- "extract text" - Extract and list all visible text
- Any other instruction - Follow it after reading the screenshot
Report results: Always mention the saved screenshot path. If analyzed, provide findings.

Example Usage

screenshot                              # Simple capture
screenshot take a look                  # Capture and analyze
screenshot is there a login button?     # Check for specific element
screenshot extract all visible text     # Extract text

Advanced Options

For advanced use cases, load references/advanced.md.

Specific simulator: Use UDID instead of booted when multiple simulators are running
Image format: --type=png (default), --type=tiff
Specific display: --display=1 (main), --display=2 (secondary, e.g. Apple Watch)
Mask sensitive data: --mask=black (iOS 15+)

Troubleshooting

"No such device: booted": No simulator running. Check with xcrun simctl list devices | grep Booted
Blank/black screenshot: Wait for content to load, ensure simulator is not minimized
Wrong simulator captured: Use specific UDID instead of booted