ai-video-generation

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install the belt CLI and related toolsets using npx skills add. These downloads originate from the inference-sh and belt-sh repositories and are fundamental to the skill's primary operation.
  • [COMMAND_EXECUTION]: The skill leverages the belt command-line utility for user authentication and triggering AI video inference models. Notably, the skill explicitly restricts the execution environment using the allowed-tools frontmatter field, limiting the agent's shell access to the belt command only.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted data by interpolating user-provided text prompts and external resource URLs into CLI arguments.
  • Ingestion points: User-provided JSON inputs passed to belt app run commands.
  • Boundary markers: None identified in the instructional examples.
  • Capability inventory: Execution of external inference models and CLI commands.
  • Sanitization: No explicit sanitization or validation of input prompts is described within the skill context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — ai-video-generation