image-to-video

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill acts as a documentation guide for the inference.sh platform. All external URLs and package installations via npx skills add point to the official repositories and documentation associated with the service, presenting no suspicious behavior.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash(belt *) tool. This permission is scoped specifically to the belt CLI, which is necessary for the skill's core functionality of running AI inference applications on the inference.sh platform.
  • [INDIRECT_PROMPT_INJECTION]: The skill templates ingest user-supplied text and file paths into shell command arguments within SKILL.md.
  • Ingestion points: User data is interpolated into the --input JSON payload for the belt command.
  • Boundary markers: None are present in the documentation examples.
  • Capability inventory: The skill uses the Bash(belt *) tool for process execution.
  • Sanitization: No specific sanitization logic is detailed in this guide.
  • Note: This is the intended primary purpose of the skill and does not represent a security vulnerability in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 11:28 AM
Security Audit — agent-trust-hub — image-to-video