remotion-captions

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install several packages from public registries. These include official Remotion packages (@remotion/captions, @remotion/whisper-webgpu), the well-known AI library @huggingface/transformers, and media utilities (mediabunny, @mediabunny/server). These are appropriate for the skill's purpose of transcribing and rendering video content.
  • [REMOTE_CODE_EXECUTION]: In display-captions.md, the skill directs the agent to fetch source code for the Basic Captions element from the official Remotion documentation website (remotion.dev) and integrate it into the project. While this involves sourcing code from an external URL, the source is the official domain of the software framework being used.
  • [DYNAMIC_EXECUTION]: The transcription logic in transcribe-captions.md involves downloading and loading AI models using downloadWhisperModel and loadWhisperModel from the @remotion/whisper-webgpu package. This typically fetches model weights from Hugging Face, which is standard practice for local AI inference.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes functionality to parse external SRT files in import-srt-captions.md and process transcription data. While processing untrusted external data (like subtitles) presents an inherent surface for indirect prompt injection, the skill uses structured parsing via @remotion/captions's parseSrt and does not provide an automated pathway for this data to influence security-critical agent decisions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 02:05 PM