fal-assets

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text to generate sprites, music, and other assets, which represents an attack surface for indirect prompt injection.
  • Ingestion points: User-supplied subjects and descriptions are passed directly to commands such as um fal sprite and um fal music in SKILL.md.
  • Boundary markers: The documentation provides examples using double quotes to encapsulate user-supplied prompt strings.
  • Capability inventory: The skill utilizes CLI tools like um, ffmpeg, and secret-tool, and performs network communication with the fal.ai API.
  • Sanitization: There are no explicit instructions for the agent to sanitize or validate user-provided strings before passing them to the execution tools.
  • [COMMAND_EXECUTION]: The skill requires the agent to execute various CLI tools including um for asset management, ffmpeg for media conversion, and pip for package installation.
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the fal Python package and communicates with official fal.ai endpoints for model inference and configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 10:07 PM