create-video-fal

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFECREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill reads sensitive credentials from ~/.gooseworks/credentials.json to authenticate requests to the media proxy.
  • [DATA_EXFILTRATION]: The gen_video.py script supports reading arbitrary local files via the @ symbol prefix in the --payload argument. This data is then sent to an external API endpoint.
  • [EXTERNAL_DOWNLOADS]: The download function in media_proxy.py retrieves files from remote URLs (e.g., FAL CDN) and writes them to the local file system.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its ingestion of external payload data. 1. Ingestion points: The --payload argument in gen_video.py. 2. Boundary markers: None identified. 3. Capability inventory: Network POST requests to a proxy and file write operations via the download function. 4. Sanitization: No sanitization or validation of the payload content is performed before transmission.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 09:42 AM
Security Audit — agent-trust-hub — create-video-fal