The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/download.py executes the gallery-dl and yt-dlp utilities via subprocess.run. The implementation correctly passes arguments as a list rather than a shell string, which effectively mitigates shell injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The skill is designed to download media files from Twitter/X. It uses external tools to perform these network operations and instructs the user to install these dependencies via package managers if they are not already available.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external URLs and fetches data from an attacker-controllable platform (Twitter/X). While it primarily handles media, the ingestion of untrusted external content constitutes a potential attack surface for indirect prompt injection.
Ingestion points: User-provided URL in scripts/download.py and the resulting media content fetched from Twitter/X.
Boundary markers: None present to delimit the input URL or the fetched content.
Capability inventory: Subprocess execution (subprocess.run), file system writes (Path.mkdir, tempfile), and directory scanning (rglob).
Sanitization: Basic URL normalization in normalize_url which handles whitespace and domain aliasing but does not perform comprehensive content validation.

twitter-media-downloader