用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/aduermael/herm --skill device-interaction命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Use calendar and location modules for user-approved Apple Calendar events and current device location.
Inspect attached or local images and visually analyze PDFs with vision. Use for photos, screenshots, scans, charts, diagrams, OCR, handwriting, UI inspection, or any task that depends on pixels, layout, or other non-structural page content.
Use the native webbrowser module for web search, browsing, authenticated website interaction, forms, file uploads and downloads, screenshots, and user handoff.
基于 SOC 职业分类
正在显示 SKILL.md
| name | device-interaction |
| description | Verify iOS app behavior on device or simulator via screenshots, UI hierarchy, and touch interactions. |
TRIGGER when: user asks to verify/test/check if the app works on device, after implementing a UI-affecting feature that needs device verification, user says "does it work", "test this", "check on device", user reports UI doesn't work as expected, need to debug touch/interaction issues. DO NOT TRIGGER when: user asks about unit tests only, build-only requests without device testing, code review without device testing, simulator configuration questions, changes that don't affect UI (e.g. comments, refactors, non-UI logic).
This is a SUBAGENT skill. Invoke it via the Agent tool when device verification is needed. If there is an open session for that work, provide that session identifier to a subagent for exclusive use by that subagent.
Agent tool:
- subagent_type: "general-purpose"
- description: "Verify login feature works"
- prompt: "Using the device-interaction skill, verify that the login feature works correctly on session <session-identifier>. Launch the app, capture screenshot and UI hierarchy, check that the login button is visible and tappable, and report if the implementation is working correctly."
After implementing a UI-affecting feature, invoke this skill to verify the implementation works on a device.
DeviceInteractionStartSession (do this early, runs in the background)
→ DeviceInteractionInstallAndRun (after each code change; includes building)
→ DeviceEventSynthesize (interact + observe, repeatable)
→ DeviceInteractionEndSession (when done — keeping sessions open is resource-heavy)
When opening a new device interaction session, pass a device identifier to select a device, or omit it to use the current destination. Pass any non-matching value to get a list of available targets.
commandLineArguments — arguments passed to the app at launch. Use $(inherited) as a token to preserve the scheme's existing arguments (e.g. ["$(inherited)", "--reset-state"] to add an extra argument at the end).environmentVariables — key/value pairs set in the app's environment at launch. Use "$(inherited)" as a key to preserve the scheme's existing environment variables (e.g. {"$(inherited)": "", "DEBUG_MODE": "1"}).Omit both parameters to leave the scheme's arguments and environment unchanged.
Prefer these parameters over editing the scheme directly. They are applied only for that one run and have no lasting effect on the user's configuration.
ALWAYS report UI issues that might be caused by code: overlapping or unreadable text, unexpectedly cropped image/text, wrong colors etc.
This tool allows performing an interaction and observing the state of a device.
The hierarchy files include calculated center positions for each element:
UIView {{100, 200}, {50, 30}}, center: {125.0, 215.0}
UIButton "Login" {{110, 205}, {30, 20}}, center: {125.0, 215.0}
{100, 200} - origin position{50, 30} - width and heightcenter: {125.0, 215.0} - calculated center point (best for tapping)Always prefer the center coordinates for touch events.
The interactionCommand parameter accepts a command syntax:
| Command | Description |
|---|---|
t <x> <y> [duration] | Tap at coordinates with optional hold duration |
d <x> <y> | Double tap |
t <x1> <y1> f <x2> <y2> [duration] | Swipe from (x1,y1) to (x2,y2) |
b h/p/u/d [duration] | Hardware button: h=Home, p=Power, u=VolUp, d=VolDown |
sender keyboard kbd <text> | Type text; must be the last command in the chain — all content after kbd is taken verbatim (multiple spaces preserved). For special characters use \u{XXXX} Unicode escapes: \u{000A} (return/newline), \u{0009} (tab) |
w duration | Wait for a duration without any work |
orientation faceDown/faceUp/landscapeLeft/landscapeRight/portrait/portraitUpsideDown | Set device orientation |
Examples:
"t 100 200" - Tap at (100, 200)"d 200 300" - Double tap at (200, 300)"t 200 600 f 200 200 0.3" - Swipe up (scroll to the content below)"t 200 200 f 200 600 0.3" - Swipe down (scroll to the content above)"b h" - Press home button"b h b h" - Press home button twice to go to the app switcher"b h w 0 b h" - Wake and unlock a device (non-passcode devices only)"sender keyboard kbd hello world" - Type text with spaces"sender keyboard kbd hello world" - Type text preserving multiple spaces"sender keyboard kbd submit\u{000A}" - Type text then press Return/submit"w 0.3" - Wait for 0.3s"orientation landscapeLeft" - Rotate device to landscapeBefore any interaction, always capture and read the hierarchy (and screenshot). After any interaction, capture again and verify the result. For complex components (like toggles or switches), look at nested elements (like Switch or Slider) — nearby elements might correspond to the actual control. When done, report findings to the main agent.
When verifying, distinguish between these categories:
print statements in the relevant code may help diagnose the issue.