Download Anything

This is a comprehensive toolkit for downloading pretty much anything from the internet. It wraps yt-dlp, aria2, gallery-dl, and spotdl with ready-made scripts for common tasks like ripping video from 1800+ sites, extracting audio, batch downloading images, or pulling Spotify playlists. The interesting part is the bilingual coverage: it includes Chinese cloud drive search engines and references for navigating both Western and Chinese resource ecosystems. The decision tree and reference docs (ebooks, software, torrents, stock media) make it practical for agents that need to locate and fetch digital resources without handholding. Honestly, the domain instability warnings are a nice touch of realism. If you're building an agent that needs to grab files from the open web, this gives you the command patterns and site directory to do it systematically.

skills.sh installs: 190
GitHub stars: 62
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add hacklyc/myagents_skills --skill download-anything --agent claude-code

Installs into .claude/skills of the current project.

Files

Download Anything

Find it. Download it. Any resource, any format.

Toolkit

# Install all tools at once
bash scripts/install-toolkit.sh

Tool	Install	Purpose
`yt-dlp`	`brew install yt-dlp`	Video/audio from 1800+ sites
`aria2c`	`brew install aria2`	Multi-thread downloads, torrents
`gallery-dl`	`pip3 install gallery-dl`	Batch image/media, 170+ sites
`spotdl`	`pip3 install spotdl`	Spotify playlists → local files
`wget`	`brew install wget`	Recursive downloads, site mirroring
`curl`	pre-installed	HTTP requests, API calls
`ffmpeg`	`brew install ffmpeg`	Media conversion
`jq`	`brew install jq`	JSON parsing for automation