Token导航 LogoToken导航

Agent Skill · 研究检索

community-ff-mcp(community FF MCP 搜索)

community-ff-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。

community-ff-mcp 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 46,181。

总安装 799周安装 32GitHub Stars 8下载 259核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:community-ff-mcp(community FF MCP 搜索)
来源仓库:https://github.com/mohn93/ff-mcp
仓库路径:skills/community-ff-mcp
安装命令:
npx skills add https://github.com/mohn93/ff-mcp --skill community-ff-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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/mohn93/ff-mcp --skill community-ff-mcp

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

community-ff-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。需要 FlutterFlow API token 和已安装的 MCP 服务器支持才能正常工作。

可通过 claude mcp add flutterflow 命令添加,并配置相应的环境变量。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。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/community-ff-mcp

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/mohn93/ff-mcp --skill community-ff-mcp;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

community-ff-mcp 怎么安装?

在终端执行 npx skills add https://github.com/mohn93/ff-mcp --skill community-ff-mcp。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

community-ff-mcp 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

community-ff-mcp 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

community-ff-mcp 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

community-ff-mcp 用的人多吗?

公开数据是累计安装 799 次,近一周安装 32 次,下载 259 次,GitHub 8 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

community-ff-mcp 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
community-ff-mcp

community FF MCP 搜索 · 本页

7998接触敏感数据未明确有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Prerequisites

This skill requires the community-ff-mcp MCP server to be installed and connected. Before proceeding, check if the list_projects tool is available. If not, the user needs to set up the MCP server first:

  1. Get a FlutterFlow API token from FlutterFlow > Profile > Account Settings > API Token (requires a paid FlutterFlow subscription)
  2. Add the MCP server to your AI client: # Claude Code claude mcp add flutterflow -e FLUTTERFLOW_API_TOKEN=<token> -- npx -y community-ff-mcp # Other clients (Claude Desktop, Cursor, Windsurf) — add to MCP config: # {"command": "npx", "args": ["-y", "community-ff-mcp"], "env": {"FLUTTERFLOW_API_TOKEN": "<token>"}}
  3. Restart your AI client, then verify by calling list_projects

Overview

The FlutterFlow MCP provides 25 tools for reading, inspecting, and editing FlutterFlow projects through YAML. It connects AI assistants to the FlutterFlow Project API, enabling programmatic access to pages, components, themes, actions, data models, and settings. All project data is represented as YAML files that can be fetched, cached locally, validated, and pushed back.

Tool Catalog

Discovery & Exploration

ToolPurpose
list_projectsList all FF projects accessible to your API token
list_project_filesList YAML file keys in a project (supports prefix filter)
list_pagesList pages with human-readable names, scaffold IDs, and folders
search_project_filesSearch file keys by keyword, prefix, or regex
sync_projectDownload all project YAML to local cache for fast reads

Reading & Understanding

ToolPurpose
get_page_by_nameFetch full page YAML by human-readable name
get_project_yamlFetch specific YAML file(s) by file key
get_page_summaryQuick page overview: widget tree, actions, params (cache-based)
get_component_summaryQuick component overview: widget tree, params (cache-based)
find_component_usagesFind all pages/components that use a given component
find_page_navigationsFind all actions that navigate to a given page

Configuration & Settings

ToolPurpose
get_themeTheme colors, typography, breakpoints, widget defaults
get_app_stateApp state variables, constants, environment settings
get_api_endpointsAPI endpoint definitions (method, URL, headers, response)
get_data_modelsData structs, enums, Firestore collections, Supabase tables
get_custom_codeCustom actions, functions, widgets, AI agents (read-only)
get_general_settingsApp Details, App Assets, Nav Bar & App Bar
get_project_setupFirebase, Languages, Platforms, Permissions, Dependencies
get_app_settingsAuthentication, Push Notifications, Deployment settings
get_in_app_purchasesStripe, Braintree, RevenueCat, Razorpay config
get_integrationsSupabase, SQLite, GitHub, Algolia, Google Maps, AdMob, etc.

Editing & Documentation

ToolPurpose
get_editing_guideGet recommended workflow and docs for a specific editing task
get_yaml_docsSearch/retrieve YAML reference docs by topic or file path
validate_yamlValidate YAML content before pushing changes
update_project_yamlPush validated YAML changes to the FF project

Core Workflows

Discover & Explore a Project

Use this workflow when first connecting to a FlutterFlow project or when you need to understand its structure.

1. list_projects → pick the target projectId
2. sync_project(projectId) → cache all YAML locally for fast reads
3. list_pages(projectId) → see all pages with human-readable names, scaffold IDs, folders
4. get_page_summary(projectId, pageName) → widget tree overview for any page of interest

After syncing, all cache-based tools (get_page_summary, get_component_summary, get_theme, get_app_state, etc.) work without additional API calls.

Read / Inspect a Page or Component

Use this workflow to understand what a page contains, how it is structured, and how it connects to the rest of the app.

1. get_page_summary(projectId, pageName) → quick overview of widget tree, actions, params
2. get_page_by_name(projectId, pageName) → full YAML if you need complete details
3. find_page_navigations(projectId, pageName) → discover what actions navigate here
4. find_component_usages(projectId, componentName) → find everywhere a component is used

For components, use get_component_summary instead of get_page_summary. Component summaries resolve nested component references so you can see the full hierarchy.

Edit an Existing Widget

Use this workflow to modify a specific widget on a page without affecting the rest of the page.

1. get_page_by_name(projectId, pageName) → get the full page YAML
2. Identify the node file key for the target widget (format: page/id-Scaffold_XXX/page-widget-tree-outline/node/id-Widget_YYY)
3. get_project_yaml(projectId, fileName: "page/id-.../node/id-Widget_XXX") → fetch the individual node YAML
4. Modify the YAML — keep inputValue and mostRecentInputValue in sync
5. validate_yaml(projectId, fileKey, content) → check for errors before pushing
6. update_project_yaml(projectId, {fileKey: content}) → push changes

Always edit at the node level. Editing the full page YAML for a single widget change risks overwriting unrelated content and is more error-prone.

Add a New Widget to a Page

Use this workflow when you need to add new widgets to an existing page.

1. get_page_by_name(projectId, pageName) → understand the current widget tree structure
2. get_yaml_docs(topic: "WidgetType") → look up the YAML schema for the widget you want to add
3. Update the page-widget-tree-outline to include a reference to the new widget key
4. Create individual node files for each new widget (one file per widget)
5. validate_yaml → validate the tree outline first, then each node file
6. update_project_yaml → push tree outline + all node files together in one call

Pushing the tree outline and node files in a single call is critical. The server strips inline widget children from the tree outline, so nodes must be separate files.

Create a Reusable Component

Use this workflow to build a new component from scratch.

1. get_yaml_docs(topic: "create component") → get the full walkthrough and required file structure
2. Design component parameters: name, dataType, isNullable for each param
3. Create these files: component metadata, widget-tree-outline, root node (with isDummyRoot: true), child nodes
4. validate_yaml → validate all files before pushing
5. update_project_yaml → push all component files in one call

Remember: the root node of a component must have isDummyRoot: true. Callback triggers use triggerType: CALLBACK with a separate parameterIdentifier field. WidgetProperty params use widgetProperty in parameterPasses, not inputValue.

Critical YAML Rules

  1. Sync inputValue and mostRecentInputValue -- Both fields must always contain the same value when you set or update them. If you change one, change both. Exceptions: fontWeightValue and fontSizeValue only accept inputValue (they have no mostRecentInputValue field).
  2. Use node-level file keys for edits -- Target page/id-Scaffold_XXX/page-widget-tree-outline/node/id-Widget_YYY for individual widget edits. Never edit the full page YAML (page/id-Scaffold_XXX) just to change a single widget. Full-page edits risk overwriting unrelated content and are harder to validate.
  3. Always validate before pushing -- Call validate_yaml before every update_project_yaml call. Validation catches missing node files, unknown fields, invalid enum values, and structural problems. Skipping validation risks corrupting the project.
  4. Push tree outline and node files together -- When adding new widgets, include the updated page-widget-tree-outline AND all individual node files in a single update_project_yaml call. Widget children embedded inline in the tree outline will be silently stripped by the FlutterFlow server.
  5. Column has no mainAxisSize field -- To achieve shrink-to-content behavior (equivalent to MainAxisSize.min in Flutter), use minSizeValue: {inputValue: true} on the Column widget instead.
  6. AppBar templateType -- Only LARGE_HEADER is a confirmed valid value. Do not use STANDARD as it may cause unexpected behavior. Control the AppBar height through the toolbarHeight property instead.
  7. Custom code is read-only -- Custom actions, functions, widgets, and AI agents cannot be created or edited through the MCP API. Use get_custom_code to read their signatures and source code, but any modifications must be made directly in the FlutterFlow UI. Attempting to push custom code changes will silently corrupt or be ignored.

Anti-Patterns

  • Don't use list_project_files to find pages -- It returns raw file keys without human-readable names. Use list_pages instead, which gives you page names, scaffold IDs, and folder assignments.
  • Don't fetch pages one-by-one to browse a project -- This is slow and wastes API calls. Use sync_project to cache everything locally first, then use get_page_summary for quick overviews of any page.
  • Don't edit full page YAML for a single widget change -- Full-page edits can overwrite other widgets, actions, or parameters. Always use node-level file keys for targeted, safe edits.
  • Don't guess YAML field names or enum values -- FlutterFlow YAML has specific field names and valid values that are not always intuitive. Use get_yaml_docs(topic) or get_editing_guide(task) to look up the correct schema before writing YAML.
  • Don't embed widget children inline in the tree outline -- The FlutterFlow server silently strips inline children from the page-widget-tree-outline file. Always create separate node files for each widget.
  • Don't push custom code changes through the API -- The API silently corrupts or ignores Dart code edits for custom actions, functions, and widgets. These must be edited in the FlutterFlow UI.

Documentation Lookup

The MCP server ships with 21 built-in reference documents. Use these tools to look up schemas, patterns, and conventions before writing YAML:

When you need...Call
Widget schema (Button, Text, Container, etc.)get_yaml_docs(topic: "Button")
Action chains, triggers, navigationget_yaml_docs(topic: "actions")
Data binding, variable sourcesget_yaml_docs(topic: "variables")
Colors, typography, dimensionsget_yaml_docs(topic: "theming")
Creating components from scratchget_yaml_docs(topic: "create component")
Editing workflows and anti-patternsget_yaml_docs(topic: "editing")
Data structs, enums, collectionsget_yaml_docs(topic: "data")
Full docs indexget_yaml_docs()
Guided workflow for a specific taskget_editing_guide(task: "change the button color")

Always consult the docs before writing YAML. They contain validated schemas, field references, enum values, and real examples.

平台分布

Codex

35.62%
按下载量换算92

Claude

26.23%
按下载量换算68

Cursor

18.57%
按下载量换算48

Gemini CLI

9.63%
按下载量换算25

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills