← Back to Skills Marketplace
Video Dl
by
dimitryvin
· GitHub ↗
· v1.1.0
1436
Downloads
0
Stars
2
Active Installs
3
Versions
Install in OpenClaw
/install video-dl
Description
Download videos from YouTube, Reddit, Twitter/X, TikTok, Instagram, and 1000+ other sites using yt-dlp. Use when user provides a video link and wants to download it.
Usage Guidance
This skill largely does what it says (downloads videos using yt-dlp), but check a few things before installing or running it:
- Ensure you have the required binaries installed from trusted sources: yt-dlp, ffmpeg and ffprobe. The skill does not declare these requirements but both scripts call them.
- Note that the SKILL.md suggests a Telegram send flow, but compress-and-send.sh only compresses and prints success; it does not actually upload or notify Telegram. Treat the CHAT_ID parameter as informational only unless you modify the script to perform uploads.
- The scripts write to ~/Downloads/videos by default and will download arbitrary URLs you pass in. Confirm you are comfortable with network downloads and legal/copyright implications for the content you request.
- Because there is no installer, nothing is written to system locations by the skill itself, but the scripts will execute external binaries and write files to your home directory — run them in a controlled environment if you are unsure.
If you plan to use this skill: verify yt-dlp and ffmpeg are installed and up-to-date, inspect or test the scripts locally, and consider running them in a sandbox or container if you are concerned about side effects.
Capability Analysis
Type: OpenClaw Skill
Name: video-dl
Version: 1.1.0
The skill bundle is classified as suspicious due to a critical shell injection vulnerability in `scripts/download.sh`. The script directly uses user-provided input for the `--output` argument in a `mkdir -p "$OUTPUT_DIR"` command without sanitization, allowing for arbitrary command execution. The `SKILL.md` further exposes this by instructing the agent to pass user-controlled options directly to the script. There is no evidence of intentional malicious behavior such as data exfiltration or backdoors.
Capability Assessment
Purpose & Capability
The skill's stated purpose (download videos via yt-dlp) matches the provided scripts. However, the SKILL.md and scripts rely on external binaries (yt-dlp, ffmpeg, ffprobe) but the skill declares no required binaries. Also SKILL.md describes a 'Sending to Telegram' workflow which implies sending to an external service, but the included compress-and-send.sh only compresses and prints status — it does not transmit to Telegram. The absence of declared binary requirements and the mismatch about Telegram are incoherent with the metadata.
Instruction Scope
SKILL.md instructs the agent/user to run the included scripts with a URL (and for compression, a CHAT_ID). The scripts operate only on the provided URL/file and write to ~/Downloads/videos (or a user-specified output). They do not read unrelated system files or additional environment variables. The scope is largely limited to downloading and compressing media, but SKILL.md's wording could mislead users into thinking the skill autonomously sends media to Telegram when it does not.
Install Mechanism
There is no install spec (instruction-only skill with two helper scripts). That minimizes install-time risk (nothing downloaded or written by an installer).
Credentials
The skill requests no environment variables or credentials (good). The scripts do access $HOME and look for $HOME/.local/bin/yt-dlp, and write to ~/Downloads/videos by default; these are reasonable for the stated task but should have been declared as required binaries/dependencies (yt-dlp, ffmpeg, ffprobe). No secrets are requested or used.
Persistence & Privilege
always is false and the skill is user-invocable only. It does not request permanent presence or attempt to modify other skills or system-wide agent settings.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install video-dl - After installation, invoke the skill by name or use
/video-dl - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.1.0
Added compress-and-send script for Telegram. Run in background to avoid chat spam.
v1.0.1
Use user-installed yt-dlp if available (fixes YouTube 403 errors on older system packages)
v1.0.0
Initial release: Download videos from YouTube, Reddit, Twitter/X, TikTok, Instagram, and 1000+ sites using yt-dlp
Metadata
Frequently Asked Questions
What is Video Dl?
Download videos from YouTube, Reddit, Twitter/X, TikTok, Instagram, and 1000+ other sites using yt-dlp. Use when user provides a video link and wants to download it. It is an AI Agent Skill for Claude Code / OpenClaw, with 1436 downloads so far.
How do I install Video Dl?
Run "/install video-dl" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Video Dl free?
Yes, Video Dl is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Video Dl support?
Video Dl is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Video Dl?
It is built and maintained by dimitryvin (@dimitryvin); the current version is v1.1.0.
More Skills