image-to-video

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides a link to installation documentation hosted on GitHub at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md. This is an external reference to instructions for the required CLI tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill demonstrates how to interpolate user-provided prompts and file paths into shell commands for AI video generation, which creates a potential surface for indirect injection.
  • Ingestion points: Prompt and image path values within the --input JSON strings in multiple examples in SKILL.md.
  • Boundary markers: The skill uses JSON formatting to structure inputs, which provides some structural separation, but does not include explicit delimiters or warnings to ignore embedded instructions in the user-supplied prompt strings.
  • Capability inventory: The skill uses Bash(belt *) to perform AI model inference tasks.
  • Sanitization: There is no evidence of sanitization or validation of the input strings before they are passed to the CLI tool.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to run commands starting with the belt keyword, as restricted by the allowed-tools frontmatter. This execution is aligned with the skill's primary purpose of using the inference.sh CLI for media generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:23 PM
Security Audit — agent-trust-hub — image-to-video