The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to generate and execute shell commands for media manipulation. It uses GNU parallel and Python's subprocess.run to orchestrate multiple ffmpeg or ImageMagick instances.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection (Category 8) where external data (such as filenames or user-provided arguments) is interpolated into command templates.
Ingestion points: Filenames gathered via find commands and user-controlled arguments like --ffmpeg-args and --out-ext in scripts/batch.py.
Boundary markers: None identified for isolating command templates from potentially malicious data in filenames.
Capability inventory: The skill has extensive command execution capabilities, including the ability to run arbitrary shell strings via GNU parallel and local scripts.
Sanitization: scripts/batch.py employs shlex.quote() for directory paths to mitigate path injection, and GNU parallel provides automatic quoting for input placeholders ({}), which reduces the risk of filename-based shell injection.

media-batch