Skip to content
Agentmart
Submit a tool

visual-control

视觉打标与精确点击控制 (Use when: 需要与 Windows GUI 应用交互、定位并点击按钮、输入框、链接等界面元素).

Skill by zzycxz

IndexedMIT★ 9
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/zzycxz-visual-control

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

视觉打标与精确点击控制 (Use when: 需要与 Windows GUI 应用交互、定位并点击按钮、输入框、链接等界面元素). NOT for: 文本处理、代码执行、非 GUI 输出解析.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/visual-control

What's inside

Skill name
visual-control
Description
视觉打标与精确点击控制 (Use when: 需要与 Windows GUI 应用交互、定位并点击按钮、输入框、链接等界面元素). NOT for: 文本处理、代码执行、非 GUI 输出解析.
License
MIT
Instructions
47 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Visual Control — 视觉打标与精确点击

基于 YOLO 的视觉感知引擎,将屏幕像素位置转化为可执行的自动化操作。

## 使用场景

✅ **以下情况使用此技能:**

- 需要点击 Windows 应用中的按钮
- 需要向表单输入框填写内容
- 动态弹窗/对话框出现,需要实时定位
- 用户说"点击"、"打开"、"找到并点击"

## 不适用场景

❌ **不使用此技能:**

- 解析文本文件内容 → 改用 `file_system_op`
- 仅需截图查看并描述 → 改用 `desktop_read_screen`

## 标准工作流

```
步骤 1:扫描并打标屏幕元素
  desktop_grounding_scan()
  → 返回带字母标签的截图(A=按钮1, B=输入框2 ...)

步骤 2:执行交互(通过宏工具 desktop_act)
  desktop_act(action="click", element_id="A", scan_cache="<上一步返回的 scan_cache JSON 字符串>")
  desktop_act(action="type", text="要输入的内容")

步骤 3:验证结果
  desktop_read_screen()  → 截图确认状态
```

## 关键原则

- **先扫描后点击**:始终先调用 `desktop_grounding_scan` 获取最新标签,再执行 `desktop_act`。
- **去噪优先**:打标时只标记交互元素(按钮、输入框、链接),过滤噪音。
- **视觉优先**:遇到动态弹窗,相信 YOLO 的实时像素结果。

## 注意事项

- 屏幕分辨率变化会影响坐标精度,高 DPI 屏幕需配置缩放补偿。
- `scan_cache` 参数必须传入上一步 `desktop_grounding_scan` 返回的完整 JSON 字符串(包含元素 id、center 坐标等信息)。不要手动构造,直接复制粘贴即可。

From zzycxz/rooster, MIT.

More from zzycxz/rooster

Signals

Stars
9
Forks
0
Open issues
0
Last commit (this folder)
Jun 9, 2026 (4 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from zzycxz

Cite this page

visual-control skill. Agent Mart. https://agentmart.dev/skills/zzycxz-visual-control (last updated 2026-10-24)