The Agent Skills Directory

[COMMAND_EXECUTION]: The skill generates and executes shell commands using the FFmpeg and FFprobe toolsets. It uses these commands to perform deep analysis of media streams, rebuild container indexes, and re-encode video/audio based on user-provided file paths or URLs.
[EXTERNAL_DOWNLOADS]: The skill can fetch and process media files from remote sources when a URL is provided as input, utilizing the network capabilities of the underlying FFmpeg installation.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its processing of untrusted external media data.
Ingestion points: User-provided file paths and URLs processed in SKILL.md and passed to diagnostic routines.
Boundary markers: No specific delimiters or "ignore instructions" warnings are applied to the input paths or the data retrieved from URLs.
Capability inventory: Extensive system-level capabilities via FFmpeg/FFprobe, including the ability to read and write files across the file system (as seen in references/diagnostics.md and references/container-repair.md).
Sanitization: The instructions do not specify validation or sanitization of input paths, leaving the agent to rely on platform-level protections against command injection.

media-fixing-and-repair