The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains extensive documentation instructing an agent to generate Python scripts that use the subprocess module to interface with ffmpeg and ffprobe. These commands are used for video decoding, audio extraction, and final output encoding. It also uses system calls like sysctl for hardware memory detection, which is consistent with the skill's performance optimization goals.- [EXTERNAL_DOWNLOADS]: The skill integrates with the ElevenLabs text-to-speech service via its public API (api.elevenlabs.io). It provides instructions for making authenticated POST requests to generate narration audio, which is a documented and well-known service functionality.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data formats, including video, audio, SRT subtitle files, and JSON layout data. This ingestion surface is documented with sanitization and parsing logic. The risk is classified as low because the data is utilized for signal processing and rendering rather than being executed as instructions within the agent's core reasoning logic.- [NO_CODE]: The skill package is composed entirely of Markdown documentation and reference guides. It does not include any pre-compiled binaries, executable scripts, or platform-specific configuration files that would run automatically upon loading.

ascii-video