Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问clear审计提醒

miro-mcpMiro MCP 搜索

Agent Skill

miro-mcp 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

5,057

周安装

215

GitHub Stars

96

下载量

1,772
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

3

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:miro-mcp(Miro MCP 搜索)
来源仓库:https://github.com/miroapp/miro-ai
仓库路径:skills/miro-mcp
安装命令:
npx skills add https://github.com/miroapp/miro-ai --skill miro-mcp
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/miroapp/miro-ai --skill miro-mcp

简介

miro-mcp 用于查找、检索和筛选相关信息。

  • 适合在关键词搜索或任务场景下快速定位候选结果。miro-mcp 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 通过 npx 命令从指定仓库安装,需参考原始 README 了解实现细节。
  • 安装前应评估是否需要联网、命令执行或文件读写权限。
  • 建议确认来源仓库的维护状态与技能的实际覆盖范围。

SKILL.md

Using Miro with Claude Code

What is Miro MCP?

Miro MCP (Model Context Protocol) enables Claude to interact directly with Miro boards. Create diagrams, documents, and tables; read board content; and extract structured documentation from visual designs.

Available Tools

Content Creation

  • diagram_get_dsl - Get the DSL format specification before creating diagrams
  • diagram_create - Generate diagrams from DSL text descriptions
  • doc_create - Create markdown documents on boards
  • table_create - Create tables with text and select columns
  • table_sync_rows - Add or update table rows

Content Reading

  • board_list_items - List items on a board with filtering by type or container
  • context_explore - Discover high-level board contents (frames, documents, prototypes, tables, diagrams)
  • context_get - Extract detailed text context from specific board items
  • table_list_rows - Read table data with column-based filtering
  • image_get_data - Get image content from boards
  • image_get_url - Get download URL for an image

Document Editing

  • doc_get - Read document content and version
  • doc_update - Edit document using find-and-replace

Board URLs and IDs

Miro tools accept board URLs directly. Extract board_id and item_id automatically from URLs like:

  • https://miro.com/app/board/uXjVK123abc=/ - Board URL
  • https://miro.com/app/board/uXjVK123abc=/?moveToWidget=3458764612345 - URL with item focus

When a URL includes moveToWidget or focusWidget parameters, the item_id is extracted automatically.

Creating Diagrams

Use diagram_create to create visual diagrams from text descriptions.

Supported Diagram Types

  • flowchart - Process flows, workflows, decision trees
  • mindmap - Hierarchical ideas, brainstorming
  • uml_class - Class structures, inheritance relationships
  • uml_sequence - Interactions between components over time
  • entity_relationship - Database schemas, data models

Description Formats

Natural language works well:

User registration flow: start -> enter email -> validate email ->
send verification -> user confirms -> create account -> redirect to dashboard

Mermaid notation for precise control:

flowchart TD
    A[Start] --> B{Valid Email?}
    B -->|Yes| C[Send Verification]
    B -->|No| D[Show Error]
    C --> E[Wait for Confirm]
    E --> F[Create Account]

Positioning Items on the Board

Board coordinates use a Cartesian system with center at (0, 0). Positive X goes right, positive Y goes down.

Spacing recommendations:

  • Diagrams: 2000-3000 units apart
  • Documents: 500-1000 units apart
  • Tables: 1500-2000 units apart

Placing in Frames

Set parent_id to a frame ID to place content inside that frame.

Creating Documents

Use doc_create to create Google Docs-style documents on boards.

Supported Markdown

  • Headings: # H1 through ###### H6
  • Bold: **text**
  • Italic: *text*
  • Unordered lists: - item
  • Ordered lists: 1. item
  • Links: [text](url)

Example Document

# Sprint Planning - Week 12

## Goals
- Complete user authentication module
- Fix critical bugs from QA

## Team Assignments
1. **Alice** - Auth backend
2. **Bob** - Frontend integration
3. **Carol** - Bug fixes

## Resources
- [Design specs](https://example.com/specs)
- [API documentation](https://example.com/api)

Working with Tables

Creating Tables

Use table_create to create tables with typed columns. Supports two column types: text for free-form entry and select for dropdowns with predefined color-coded options.

Adding and Updating Rows

Use table_sync_rows to add or update table data. Set key_column to match existing rows for upsert behavior — matching rows are updated, non-matching rows are inserted as new.

Reading Table Data

Use table_list_rows to read table contents. Filter by column value using ColumnName=Value format.

Extracting Board Content

Discovering Board Contents

Use context_explore to get a high-level view of what's on a board — returns frames, documents, prototypes, tables, and diagrams with their URLs and titles.

Getting Item Details

Use context_get to extract detailed content from specific items:

Item TypeReturns
DocumentsHTML markup of the document content
Prototype screensHTML markup representing the UI/layout
Prototype containersAI-generated map of all screens with navigation flow
FramesAI-generated summary of frame contents
TablesFormatted table data
DiagramsAI-generated description and analysis

Workflow

  1. Call context_explore to discover board contents
  2. Identify items of interest from the results
  3. Call context_get with specific item URLs (with moveToWidget parameter)

Browsing Board Items

Use board_list_items to explore board contents. Filter by item type (frame, sticky_note, card, shape, text, image, document, embed) or by container to list items within a specific frame.

Best Practices

For Diagrams

  • Be specific about elements and relationships
  • Specify flow direction (top-down, left-right)
  • Include decision points and conditions
  • Let AI auto-detect diagram type or specify explicitly

For Documents

  • Structure with clear headings
  • Keep content focused and scannable
  • Use lists for multiple items
  • Include links to related resources

For Tables

  • Choose meaningful column names
  • Use select columns for status/priority fields
  • Define clear, distinct option colors
  • Use key_column for idempotent updates

For Context Extraction

  • Start with context_explore to discover board contents
  • Focus on specific frames when boards are large
  • Use context_get with item URLs for detailed content

Quick Reference

TaskTool
Get diagram DSL specdiagram_get_dsl
Create diagramdiagram_create
Create documentdoc_create
Read documentdoc_get
Edit documentdoc_update
Create tabletable_create
Add/update table rowstable_sync_rows
Read table datatable_list_rows
Discover board contentscontext_explore
Extract item detailscontext_get
List board itemsboard_list_items
Get image dataimage_get_data
Get image URLimage_get_url

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude Code

31.21%
按下载量换算553

windsurf

24.35%
按下载量换算431

trae

18.18%
按下载量换算322

OpenCode

13.3%
按下载量换算236

Cursor

7.72%
按下载量换算137

kiro-cli

3.96%
按下载量换算70

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。

来源信息

继续浏览同类 Skills