threejs-frame-animation
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests and processes external video/image assets and structured manifest JSON data.
- Ingestion points: Video files are ingested in 'scripts/video-to-atlas.mjs', while image files and 'source-manifest' JSON files are processed in 'scripts/images-to-atlas.mjs' and 'scripts/image-frame-source.mjs'.
- Boundary markers: Instructions explicitly require the agent to propose a design concept and wait for human confirmation before generating assets. Manifests are validated against the 'threejs-image-frame-source/1' schema.
- Capability inventory: The skill utilizes filesystem read/write access (via 'node:fs/promises') and executes external commands ('ffmpeg' and 'ffprobe' via 'node:child_process').
- Sanitization: Mitigation measures include HTML escaping in 'scripts/frame-atlas-core.mjs' for the preview generator and strict path validation in 'scripts/image-frame-source.mjs' (using 'realpath' and 'isWithinRoot') to prevent directory traversal attacks via manifest file paths.
- [COMMAND_EXECUTION]: The script 'scripts/video-to-atlas.mjs' executes 'ffmpeg' and 'ffprobe' using 'node:child_process.spawn'. The use of argument arrays rather than shell strings effectively prevents shell injection. This capability is localized and necessary for the skill's primary function of media processing.
Audit Metadata