Skills
Navigate
skills/halt-catch-fire/skills/image-to-video/Gen Agent Trust Hub

image-to-video

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute belt CLI commands for inference and media processing, which is limited to the tool's intended functionality via platform restrictions.
  • [EXTERNAL_DOWNLOADS]: The skill directs users to install the belt-sh/cli package and references installation documentation hosted on GitHub (inference-sh/skills).
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it ingests user-provided text for prompts and interpolates it into CLI command arguments.
  • Ingestion points: User-defined strings for the prompt key in JSON payloads passed to belt app run commands.
  • Boundary markers: User input is encapsulated within JSON objects provided as command-line arguments.
  • Capability inventory: The skill is constrained to using the belt command via the Bash tool.
  • Sanitization: The skill does not implement explicit sanitization or filtering for user-provided prompts within the provided examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 03:53 PM
Security Audit — agent-trust-hub — image-to-video