fal-text-to-image

Pass

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides Python scripts (fal-text-to-image, fal-image-remix, fal-image-edit) intended for execution by the agent to interface with image generation models.\n- [EXTERNAL_DOWNLOADS]: Downloads generated image assets from fal.ai's official domains to the local outputs directory.\n- [DATA_EXFILTRATION]: Transfers user-selected local images to the fal.ai processing service. This network operation is restricted to the intended purpose of image transformation and editing.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface where external data (user prompts) is interpolated into API requests without sanitization or boundary markers.\n
  • Ingestion points: Prompt arguments in fal-text-to-image, fal-image-remix, and fal-image-edit.\n
  • Boundary markers: Absent; user input is passed as a raw string to the external model.\n
  • Capability inventory: Network communication via fal-client and requests; local file system writes via PIL.\n
  • Sanitization: No validation, escaping, or filtering of the input prompt is performed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 31, 2026, 11:51 AM