The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous examples of shell commands using ffmpeg and ffprobe for media transformation tasks. This is the core functionality of the skill and is intended for legitimate video production workflows.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates processing external media files (video, audio, GIFs) and text-based file lists. If these files contain malicious metadata or crafted content, they could potentially influence the agent's behavior when it parses the output of the processing tools.
Ingestion points: Input media files and list.txt files are read by ffmpeg and ffprobe commands in SKILL.md and reference.md.
Boundary markers: No specific delimiters or instructions are provided to the agent to treat media metadata or command outputs as untrusted data.
Capability inventory: The skill facilitates the execution of ffmpeg and ffprobe commands and the creation of local text files via shell redirection.
Sanitization: There is no evidence of sanitization or validation of input file names or internal metadata before they are processed by the shell commands.

ffmpeg