The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads necessary assets and source code from established repositories and official domains.
Model weights for transcription are downloaded using the downloadWhisperModel function from @remotion/whisper-webgpu.
Component source code is fetched from the official documentation at https://www.remotion.dev/elements/captions/basic-captions.md.
[COMMAND_EXECUTION]: Employs standard project management commands to handle dependencies.
The skill provides instructions to use npx remotion add, bunx remotion add, yarn remotion add, and pnpm exec remotion add to install the required @remotion/captions and @remotion/whisper-webgpu packages.
[INDIRECT_PROMPT_INJECTION]: The skill processes external media and text files, which represents a standard architectural attack surface for indirect prompt injection.
Ingestion points: Ingests user-provided video/audio files for transcription (transcribe-captions.md) and .srt files for parsing (import-srt-captions.md).
Boundary markers: No specific boundary markers or "ignore instructions" warnings are applied to the content extracted from these files.
Capability inventory: The skill includes capabilities to write to the file system (writeFile in transcribe-captions.md) and make network requests (fetch in import-srt-captions.md).
Sanitization: The skill transcribes and parses data literally without content sanitization.
[DATA_EXFILTRATION]: Network operations are restricted to legitimate asset retrieval.
The skill supports fetching remote .srt files via fetch(), which is a documented feature for importing subtitles and does not involve access to sensitive system files.

remotion-captions