skills/inference-sh/skills/p-video/Gen Agent Trust Hub

p-video

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references installation instructions for the belt CLI tool hosted on the official GitHub repository for inference-sh. This is a legitimate dependency from the vendor.
  • [COMMAND_EXECUTION]: The instructions utilize the belt command-line tool to interface with Pruna video models. The skill's configuration correctly limits the agent's environment to this specific tool.
  • [PROMPT_INJECTION]: The skill manages an indirect prompt injection surface through the ingestion of external data:
  • Ingestion points: User-supplied prompt, image, and audio parameters passed to the belt app run command in SKILL.md.
  • Boundary markers: Absent; there are no specific delimiters defined to separate user input from the agent's instructions.
  • Capability inventory: The skill uses the belt command to transmit data to remote video generation services.
  • Sanitization: No explicit sanitization or input validation is described for the user-provided prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 09:34 PM