Social Video Downloader
Analysis
This is a coherent social-video downloader with expected local command and package-install risks, but no evidence of hidden credential use, exfiltration, persistence, or destructive behavior.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
pip install --break-system-packages yt-dlp # macOS brew install yt-dlp ffmpeg
The setup asks users to install external command-line dependencies without version pinning; this is normal for a yt-dlp-based downloader, but it relies on trusted package sources and modifies the local environment.
cmd = [
"yt-dlp",
"-o", output_file,
"--no-playlist",
...
"--", url
]
result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)The skill invokes the local yt-dlp binary through subprocess to download media. This is central to the stated purpose and the command is constructed without shell interpolation.
meta_cmd = ["yt-dlp", "--no-playlist", "--print", "title", "--no-warnings", "--", url]
...
print(f"Title: {meta_result.stdout.strip()}")The script prints a video title fetched from the external platform into the agent-visible output. That content should be treated as untrusted text and not as instructions or as a status marker.
