The Agent Skills Directory

[DATA_EXPOSURE]: The script assets/downloader.py contains a hardcoded public API key (ak_b45b6a6881f4400fb321428947eb6661). While this is a vendor-provided key for trial use, hardcoding credentials is a suboptimal security practice as it facilitates unauthorized harvesting.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided URLs at runtime, creating an attack surface for indirect prompt injection. 1. Ingestion points: The URL parameter passed to the downloader.py script via command-line arguments. 2. Boundary markers: The skill does not implement explicit delimiters or instruction guards to isolate user-provided content. 3. Capability inventory: The skill possesses network access (via the requests library) and the ability to write files to the local filesystem. 4. Sanitization: The script includes a sanitize_filename function to mitigate directory traversal and utilizes regular expressions to validate that incoming URLs originate from supported domains.
[REMOTE_CODE_EXECUTION]: The skill invokes a local Python script using shell commands. If the AI agent fails to properly escape the user-provided URL before passing it to the command-line tool, it could potentially be exploited for command injection.

video-downloader