Fabric Mcp

io.github.mpzarde由 GitHub 账号 mpzarde 发布v1.0.0更新于 Oct 24, 2026

MCP server for Fabric AI patterns

发布方自己的服务:Fabric

概览

AI 生成的概览

让助手对文本、文件、网址和 YouTube 字幕运行 Fabric AI 模式,用于摘要、分析和学习。

可以做什么
封装 Fabric CLI,使助手可以列出可用模式、按名称运行任意模式,并使用预配置工具,例如 extract_wisdom、summarize、analyze_claims、create_quiz、to_flashcards、explain_code、summarize_git_diff、analyze_logs 和 analyze_incident。组合工具可一步完成抓取内容并应用模式:analyze_youtube_video、analyze_file 和 analyze_url。health_check 工具用于检查 Fabric 与 yt-dlp 是否已安装并配置。
适用场景
适合希望在 Claude Desktop 或 Warp 中直接使用 Fabric 模式库的场景,例如总结文章、从视频中提取要点、生成闪卡或测验、审阅 git diff,以及分析日志和安全事件。如果你已经在使用 Fabric,并希望不离开助手就能调用其模式,这个服务器很合适。
运行要求
以 Node.js 本地进程通过 stdio 运行,仅支持桌面端。需要安装 Fabric CLI 并在 PATH 中可用;YouTube 字幕功能还需要 yt-dlp。可选环境变量 FABRIC_PATH 和 YTDLP_PATH 用于指定自定义 CLI 路径。Fabric 本身需要配置 AI 提供商和 API 密钥。
安装前请注意
该服务器调用 Fabric CLI,会把传入的内容发送到 Fabric 所配置的 AI 提供商,因此除非该提供商可接受,否则不要提交敏感材料。运行模式会消耗 Fabric 提供商的配额,输入过大时可能超时。它会在被要求时读取本地文件并抓取网址,请只指向你愿意分享的内容。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Fabric Mcp,将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。

其他 MCP 客户端

参照 仓库 中的启动说明。

README

Fabric MCP Server

Model Context Protocol (MCP) server for Fabric - Daniel Miessler's AI-powered content processing framework.

This MCP server allows Claude (via Claude Desktop or Warp CLI) to use Fabric's extensive collection of AI patterns for content analysis, summarization, learning, coding, security analysis, and more.

Quick Start

  1. Install Fabric (if not already installed):

    See the official Fabric installation guide for detailed instructions.

  2. Install yt-dlp (optional, for YouTube transcripts):

    bash
    # macOSbrew install yt-dlp
    # Linux/Windowspip install yt-dlp
  3. Download and Install:

    • Download fabric-mcp.mcpb from Releases
    • Double-click the file
    • Click "Install" in Claude Desktop
    • Restart Claude Desktop
  4. Try it: Ask Claude: "List all fabric patterns"

Features

  • 🚀 Direct CLI Integration: Uses Fabric CLI directly (no REST API server needed)
  • 🔧 19 Pre-configured Tools: Direct access to commonly-used Fabric patterns plus combined analysis tools
  • 🎯 Generic Pattern Runner: Run any Fabric pattern with run_fabric_pattern
  • 📺 YouTube Support: Fetch video transcripts directly (requires yt-dlp)
  • 🔍 Pattern Discovery: List all available patterns dynamically
  • 📁 File & URL Analysis: Process files and web content with Fabric patterns
  • 🏥 Health Check: Verify dependencies and configuration

Prerequisites

1. Install Fabric

If you don't have Fabric installed, follow the official Fabric installation guide.

2. Install yt-dlp (Optional)

Required only if you want to fetch YouTube transcripts:

bash
# macOSbrew install yt-dlp
# Linuxpip install yt-dlp# Or: sudo apt install yt-dlp (on Ubuntu/Debian)
# Windowspip install yt-dlp

3. Verify Installation

bash
# Test Fabricfabric --versionfabric --listpatterns
# Test yt-dlp (optional)yt-dlp --version

Installation

Option 1: Download and Install (Recommended) 🎉

The easiest way to install:

  1. Download the latest fabric-mcp.mcpb from Releases
  2. Double-click the downloaded file
  3. Claude Desktop will open and prompt you to install
  4. Click Install
  5. Restart Claude Desktop
  6. Done!

The .mcpb file is a self-contained bundle that includes everything needed.

Option 2: Build from Source

If you want to build from source:

bash
git clone https://github.com/mpzarde/fabric-mcp.gitcd fabric-mcpnpm installnpm run build

Then either:

  • Build the .mcpb bundle: ./build-mcpb.sh (creates fabric-mcp.mcpb you can install)
  • Configure manually: See Configuration section below for manual setup

Configuration

For Claude Desktop

Note: If you installed using the .mcpb bundle (Option 1), Claude Desktop configures this automatically. This section is only needed for manual builds.

Add to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%/Claude/claude_desktop_config.json

json
{  "mcpServers": {    "fabric": {      "command": "node",      "args": ["/absolute/path/to/fabric-mcp/dist/index.js"]    }  }}

Replace /absolute/path/to/fabric-mcp/ with your actual project path.

Optional environment variables (if needed):

json
"env": {  "FABRIC_PATH": "/custom/path/to/fabric",  "YTDLP_PATH": "/custom/path/to/yt-dlp"}

For Warp CLI

Note: Manual configuration is required for Warp. Build from source or extract the .mcpb bundle first.

Add to your Warp MCP configuration file:

macOS/Linux: ~/.warp/mcp_config.json

json
{  "mcpServers": {    "fabric": {      "command": "node",      "args": ["/absolute/path/to/fabric-mcp/dist/index.js"]    }  }}

Replace /absolute/path/to/fabric-mcp/ with your actual project path.

Optional environment variables (if needed):

json
"env": {  "FABRIC_PATH": "/custom/path/to/fabric",  "YTDLP_PATH": "/custom/path/to/yt-dlp"}

Restart Claude/Warp

After configuration:

  • Claude Desktop: Restart the application
  • Warp: Restart the terminal or run warp mcp reload

Available Tools

Core Tools

  1. health_check - Check if Fabric, yt-dlp, and other dependencies are installed and configured
  2. list_fabric_patterns - List all available Fabric patterns
  3. run_fabric_pattern - Run any Fabric pattern with custom input
  4. get_youtube_transcript - Fetch transcript from YouTube video

Combined Analysis Tools

These tools streamline common workflows by fetching content and applying patterns in one step. Benefits:

  • ✅ No truncation: Complete content is passed directly to Fabric
  • ✅ More efficient: Single tool call instead of multiple
  • ✅ Better results: Fabric processes the full content without context window limits
  1. analyze_youtube_video - Fetch YouTube transcript and apply a pattern in one step
  2. analyze_file - Read a file and apply a pattern in one step
  3. analyze_url - Fetch URL content and apply a pattern in one step

Pre-configured Pattern Tools

These patterns are exposed as dedicated tools with the fabric_ prefix:

Content Analysis
  • fabric_extract_wisdom - Extract key insights and quotes from content
  • fabric_summarize - Create concise summaries
  • fabric_analyze_claims - Analyze and fact-check claims
  • fabric_analyze_paper - Analyze academic papers
Learning
  • fabric_create_quiz - Generate quiz questions
  • fabric_to_flashcards - Convert content to flashcards
Development
  • fabric_summarize_git_diff - Summarize git diffs for reviews
  • fabric_create_coding_project - Generate project structure from ideas
  • fabric_explain_code - Explain code in simple terms
Security & Operations
  • fabric_analyze_logs - Analyze log files
  • fabric_analyze_incident - Analyze security incidents
Writing
  • fabric_improve_writing - Improve writing quality

Usage Examples

With Claude Desktop

Extract wisdom from a YouTube video:

User: Can you extract the key insights from this video?       https://www.youtube.com/watch?v=dQw4w9WgXcQ
Claude: [Uses get_youtube_transcript + fabric_extract_wisdom]

Summarize an article:

User: Summarize this article: [paste article text]
Claude: [Uses fabric_summarize]

Create flashcards from lecture notes:

User: Convert these lecture notes into flashcards: [paste notes]
Claude: [Uses fabric_to_flashcards]

Analyze a git diff:

User: Review this git diff and summarize the changes:      [paste git diff output]
Claude: [Uses fabric_summarize_git_diff]

Combined Analysis Tools

For efficiency, use the combined tools that fetch and analyze in one step:

User: Extract wisdom from this YouTube video:      https://www.youtube.com/watch?v=example
Claude: [Uses analyze_youtube_video with pattern='extract_wisdom']
User: Summarize this article:      https://example.com/article
Claude: [Uses analyze_url with pattern='summarize']
User: Explain the code in /path/to/script.py
Claude: [Uses analyze_file with pattern='explain_code']

Chaining Patterns

Claude can chain multiple patterns together:

User: Get the transcript from this video, extract the wisdom,       then create flashcards from it.      https://www.youtube.com/watch?v=example
Claude: [1. Uses get_youtube_transcript][2. Uses fabric_extract_wisdom on transcript][3. Uses fabric_to_flashcards on extracted wisdom]

Use Cases

Based on Daniel Miessler's Fabric patterns:

Content Processing

  • YouTube video → extract wisdom / summarize
  • Article / blog post → summarize, analyze claims
  • Podcast transcript → extract wisdom, pull quotes

Writing & Development

  • Idea → essay (via create_coding_project or similar patterns)
  • Code diff → PR description (via summarize_git_diff)
  • Meeting notes → user stories (via run_fabric_pattern with agility_story)

Learning

  • Any content → flashcards (via to_flashcards)
  • Dense docs → glossary with analogies
  • Video/article → quiz (via create_quiz)

Security Analysis

  • Malware samples → analysis (via analyze_malware)
  • Log files → threat interpretation (via analyze_logs)
  • Security scans → rules (via run_fabric_pattern)

Troubleshooting

Check server health

Ask Claude: "Run a health check" - this will verify:

  • Fabric installation and version
  • Fabric configuration (API keys, patterns)
  • yt-dlp installation (for YouTube support)
  • Provide installation instructions if anything is missing

Common Issues

Fabric not found

Error: Failed to spawn fabric: ENOENT

Solutions:

  1. Install Fabric following the official installation guide
  2. Verify installation: which fabric and fabric --version
  3. If using a custom Fabric location, set FABRIC_PATH in your MCP config
Pattern not found

Error: Pattern not found: pattern_name

Solutions:

  1. Ask Claude to list all patterns: "List all fabric patterns"
  2. Update patterns: fabric --updatepatterns
  3. Check spelling - use underscores (e.g., extract_wisdom not extractWisdom)
YouTube transcripts not working

Error: Failed to fetch YouTube transcript

Solutions:

  1. Install yt-dlp: brew install yt-dlp (macOS) or pip install yt-dlp
  2. Verify: yt-dlp --version
  3. If using custom location, set YTDLP_PATH in your MCP config
MCP server not appearing in Claude

Solutions:

  1. Verify config file location:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%/Claude/claude_desktop_config.json
  2. Check JSON syntax with a validator
  3. Ensure path to dist/index.js is absolute
  4. Restart Claude Desktop completely
  5. Check logs: ~/Library/Logs/Claude/mcp-server-fabric.log
Pattern execution timeout

Error: Fabric command timed out

Solutions:

  • Large patterns may take time (default: 5 minutes)
  • Check your Fabric API keys and quota
  • Try with a smaller input first

Development

Build

bash
# Build TypeScript onlynpm run build
# Build complete .mcpb bundle./build-mcpb.sh

Watch mode

bash
npm run watch

Testing locally

bash
# Build the projectnpm run build
# Test the MCP servernode dist/index.js# Server will run on stdio and wait for MCP protocol messages
# Or use the health check to verify dependenciesecho '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"health_check","arguments":{}}}' | node dist/index.js

Architecture

┌─────────────────┐│  Claude/Warp    │└────────┬────────┘         │ MCP Protocol (stdio)         ▼┌─────────────────┐│  MCP Server     │ ← This project│  (Node.js)      │└────────┬────────┘         │ Direct CLI calls         ▼┌─────────────────┐│  Fabric CLI     ││  (fabric cmd)   │└────────┬────────┘         │         ▼┌─────────────────┐│  AI Patterns    ││  OpenAI/etc.    │└─────────────────┘

The MCP server:

  1. Exposes Fabric patterns as MCP tools
  2. Calls Fabric CLI directly with pattern names and input
  3. Returns results back to Claude/Warp
  4. No REST API server needed - uses CLI interface

Resources

License

MIT

Contributing

Contributions welcome! Please open an issue or PR.

Credits

来源:README.md,提交 dc3238e

工具

0
SourceWeft 尚未在沙箱中启动过该软件包,因此其工具未知。

版本历史

1
  1. v1.0.0最新Sep 16, 2026

评分

已满足适用检查项的 50%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通25 分

    请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。

  • 已声明认证方式10 分

    请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。

  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分

其他 Fabric 集成