據 AlphaMoat 統計,截至 2026-09,Media Downloader 累計下載 4,428 次、獲 5 個星標,在全部 5,078 個 AI 技能中排名第 2,197 位,在「設計創意」分類(512 個)中排名第 214 位。
| 平台 | 累計安裝 | 累計下載 | 星標 | 評論 |
|---|---|---|---|---|
| ClawHub ↗ | 137 | 4211 | 5 | - |
| SkillHub ↗ | 437 | 4428 | 5 | — |
各平台的計數口徑不同,統計時間不一致。
指令碼透過uv執行,按媒體型別儲存到影片或音樂資料夾,並返回檔案或列表目錄路徑。可指定輸出目錄和可選cookies檔案;Telegram音訊傳送屬於單獨的交付分支。 下載會占用磁碟空間,檔案也可能由區域網路媒體伺服器分享。請使用專用資料夾並保護 cookies 檔案。本版本另設播放清單數量與檔案大小上限;選用 Telegram 傳送前先核對聊天對象與檔案。
資料釋出於 2026-10-27
以下是根據技能文件生成的示例,不代表實際執行結果。
2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。
| 月份 | 累計下載 | 環比 | 星標 |
|---|---|---|---|
| 2026-09 | 4428 | — | 5 |
| 2026-08 | 4046 | +4.7% | 5 |
| 2026-07 | 3864 | +7.3% | 5 |
| 2026-06 | 3600 | +3.0% | 5 |
| 2026-05 | 3495 | — | 5 |
| 2026-03 | 3163 | — | 5 |
以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。
---
name: dl
description: Download Video/Music from YouTube/Bilibili/X/etc.
author: guoqiao
metadata: {"openclaw":{"always":false,"emoji":"🦞","homepage":"https://clawhub.ai/guoqiao/dl","os":["darwin","linux","win32"],"requires":{"bins":["uv"]}}}
triggers:
- "/dl <url>"
- "Download this video ..."
- "Download this music ..."
---
# Media Downloader
Smartly download media (Video/Music) from URLs (YouTube, Bilibili, X, etc.) to the appropriate local folders.
- **Video:** Save into `~/Movies/` or `~/Videos/`.
- **Music:** Save into `~/Music/`.
- **Playlists:** Saves items into a subdirectory (e.g., `~/Music/<playlist_name>/`).
Designed to work with a local Media Server (e.g., Universal Media Server, Jellyfin) for instant playback on TV/devices.
## Agent Procedure
When the user provides a URL or asks to download media, **you MUST follow this exact sequence:**
1. **Acknowledge:**
- Immediately reply to the user: "Downloading with dl skill..."
2. **Execute:**
- Run the script:
```bash
uv run --script ${baseDir}/dl.py "<url>"
```
3. **Capture Path:**
- Read the script output, a path will be printed to stdout at the end, points to either a single file or a folder contains the playlist items.
4. **Upload (Telegram Only):**
- If the user is on Telegram (check context or session) AND the file is audio (mp3/m4a):
- Use the `message` tool to send the file to the user:
```json
{
"action": "send",
"filePath": "<filepath>",
"caption": "Here is your music."
}
```
## Usage
Run `dl.py` as a uv script:
```bash
# save into default dirs ~/Music or ~/Movies or ~/Videos
uv run --script ${baseDir}/dl.py <url>
# specify your own output dir
uv run --script ${baseDir}/dl.py <url> -o <out_dir>
```
The script will print output path, either a file or a folder.
A optional cookies file could be provided to make yt-dlp more reliable, with which ever of these detected first:
- `${baseDir}/.cookies.txt`
- `$DL_COOKIES_FILE`
- `$COOKIES_FILE`
- `~/.cookies.txt`
## Setup (User)
This skill will be much more useful if you setup a media server on same machine to share the downloaded media in your LAN:
1. Install a DLNA/UPnP Media Server (Universal Media Server, miniDLNA, Jellyfin).
2. Share `~/Music` and `~/Movies` (or `~/Videos`) folders.
3. Downloaded media will appear automatically on your TV, with apps support DLNA/UPnP, such as VLC.
See [example script](https://github.com/guoqiao/skills/blob/main/dl/ums/ums_install.sh) to setup Universal Media Server on Mac.
ContextWeave 架構圖一鍵生成
用 Nano Banana Pro 生成或編輯圖片的指令碼技能。
用 whisper 命令列工具在本機將音訊轉成文字或進行翻譯。
透過 MCP 服務生成與編輯圖表的技能。
面向介面設計的思路與落地建議技能,覆蓋佈局、互動流程、設計令牌與前端改動。
面向多家影片模型 API 的 AI 影片生成與編排技能。
用 ffmpeg 從影片裡取出畫面幀,或生成便於快速檢視的縮圖。
用本地 Whisper 把音訊檔案轉成文字,模型下載完成後可完全離線執行。
資料月份:2026-09 · 下載/星標/安裝數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。