kajisho5/ffmpeg-skill

❌ 实测失败serverPythonMIT★ 4948

Local video editor for coding agents: 42 FFmpeg tools with a machine-readable contract and an MCP server, for Claude Code, Cursor and Codex. Cut, captions, reframe, loudness, platform exports. No cloud, no API keys.

实测结果

失败

沙箱内启动/握手失败，下方给出失败阶段与原始报错。失败阶段：initialize，错误类型：missing_system_dep。这类信息对判断“是不是我这边的环境问题”很有用，原始报错见下。

error_type: missing_system_dep

warning: ffmpeg was not found on PATH. The skill needs FFmpeg to run:
    macOS:   brew install ffmpeg-full; export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"
             # keg-only, so installing it does not put it on PATH: keep the export in your shell profile;
             # the plain ffmpeg formula lacks subtitles/drawtext/zscale
    Ubuntu:  sudo apt install ffmpeg
    Windows: winget install Gyan.FFmpeg


  warning: no Python 3.10+ interpreter was found (tried: python3); the skill's scripts need one (3.0.0 removed Python 3.9).

安装方式: npx
包名: ffmpeg-skill
解析依据: npm registry match (has bin)
沙箱: docker
总耗时: 5162 ms
安装耗时: 3935 ms
工作目录: 3 MB
探针版本: 1.0.0
实测时间: 2026-11-11 07:50:41

如何使用

❌ 实测失败missing_system_dep · warning: ffmpeg was not found on PATH. The skill needs FFmpeg to run；下面配置仅供参考，跑不跑得通取决于你的环境

1安装 / 启动命令npm / npx

npx -y ffmpeg-skill

2客户端配置Claude Desktop / Cursor

{
  "mcpServers": {
    "kajisho5-ffmpeg-skill": {
      "command": "npx",
      "args": [
        "-y",
        "ffmpeg-skill"
      ]
    }
  }
}

整段粘进 claude_desktop_config.json 的 mcpServers 里（Cursor 用 ~/.cursor/mcp.json），保存后重启客户端。

安装方式来源：npm registry match (has bin)，沙箱实测未通过。

打开 GitHub 仓库 ↗项目主页 ↗MCP 体检工具

仓库信息

全名: kajisho5/ffmpeg-skill
分类: server
Stars: 4948
Forks: 407
主要语言: Python
License: MIT
最近推送: 2026-11-10
仓库创建: 2026-09-03
首次收录: 2026-11-10
最近抓取: 2026-11-10

#agent-skills#ai-agents#audio-processing#captions#claude-code#codex#coding-agent#cursor#ffmpeg#loudness-normalization#mcp#mcp-server