The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to clone a repository from an unverified GitHub account (github.com/louisedesadeleer/clipify.git).
[REMOTE_CODE_EXECUTION]: The installation process involves downloading external scripts from an unverified source and subsequently executing them as part of the skill's primary workflow. This pattern is a significant risk for remote code execution if the source repository were to be compromised or malicious.
[COMMAND_EXECUTION]: The skill heavily utilizes shell commands and Python's subprocess.run to interact with the filesystem and external tools like FFmpeg.
[PROMPT_INJECTION]: Indirect prompt injection surface exists where untrusted data (video content transcribed via Whisper) is processed by the skill to generate subtitles and titles.
Ingestion points: Whisper transcription JSON files (source.json).
Boundary markers: None identified in the provided script examples.
Capability inventory: File system writes and shell command execution via subprocess.run (found in batch processing and integration sections).
Sanitization: No explicit sanitization of transcribed text is shown before it is passed to FFmpeg filters or other scripts.

clipify-video-clip-generator