The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses a Python script, scripts/generate_bundle.py, to dynamically create bash scripts (.sh) for FFmpeg processing. These scripts are constructed by inserting file paths and other parameters into templates using string formatting. Without proper shell-escaping, this pattern is susceptible to command injection if the inputs contain shell metacharacters.
[PRIVILEGE_ESCALATION]: The scripts/generate_bundle.py script automatically changes the file permissions of the generated shell scripts to make them executable by calling out_path.chmod(0o755).
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of Whisper JSON or SRT transcript files.
Ingestion points: The skill reads external transcript files (Whisper JSON, SRT) as inputs to identify chapters and clips.
Boundary markers: There are no explicit delimiters used to separate the transcript data from the generated instructions.
Capability inventory: The skill can perform file writes and generate executable shell scripts that run external tools like FFmpeg.
Sanitization: Input data is not sanitized for potential injection characters before being processed.
[REMOTE_CODE_EXECUTION]: The skill includes instructions to install the python-bidi package via pip and the ffmpeg tool via system package managers to support Hebrew text processing and video editing.

hebrew-podcast-postproduction