fal-assets
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of user-supplied text prompts and local reference files (e.g., using the
--refflag in theum fal editcommand). This creates a surface where malicious instructions could be embedded in the input data to influence the AI's generation or behavior. - Ingestion points: User-provided prompt strings and reference image/3D files in
SKILL.mdcommands. - Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the input data.
- Capability inventory: The skill performs network operations to external APIs (fal.ai) and executes local utility commands (
um,ffmpeg). - Sanitization: There is no explicit sanitization or validation of the input content before processing.
- [EXTERNAL_DOWNLOADS]: The skill interacts with and downloads media assets from well-known technology providers.
- Downloads generated assets from
fal.aihosted infrastructure. - References and utilizes endpoints from established services including OpenAI, ElevenLabs, ByteDance, and Google.
Audit Metadata