remotion-best-practices
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests documentation from remotion.dev and media metadata from external URLs, which could host untrusted instructions.
- Ingestion points: Markdown content fetched from remotion.dev and lottiefiles.com, and data retrieved from user-provided URLs in metadata functions.
- Boundary markers: Examples do not utilize specific delimiters to isolate external text from the agent's core instructions.
- Capability inventory: Subprocess execution is supported through npx remotion CLI commands for project management and rendering.
- Sanitization: No explicit filtering of fetched documentation or data is described in the best-practice examples.
- [CREDENTIALS_UNSAFE]: Instructions for documentation search include a hardcoded public API key.
- File: remotion-docs/REFERENCE.md
- Evidence: x-algolia-api-key=3e42dbd4f895fe93ff5cf40d860c4a85
- Context: This key is used for searching public Remotion documentation and is intended to be available in frontend search interfaces.
- [EXTERNAL_DOWNLOADS]: Fetches data and assets from established vendor and service domains.
- Operations include retrieving documentation from remotion.dev and querying algolia.net.
- Media assets are referenced from vendor-controlled domains like remotion.media.
- [COMMAND_EXECUTION]: The skill frequently uses the framework's CLI for development tasks.
- Commands like npx remotion studio, npx remotion render, and npx remotion add are used for project operations and dependency management.
Audit Metadata