Skills
skills/inferencesh/skills/video-ad-specs/Gen Agent Trust Hub

video-ad-specs

Fail

Audited by Gen Agent Trust Hub on Mar 22, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides a command to fetch an installation script from the vendor's domain (https://cli.inference.sh) and executes it by piping the output directly to the shell (| sh). This is the documented method for installing the vendor's CLI tool.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute infsh commands for user authentication and running media generation applications.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it interpolates user-provided text into JSON payloads passed to remote media generation applications.
  • Ingestion points: User prompts in SKILL.md are passed to apps like google/veo-3-1-fast and bytedance/seedance-1-5-pro via the --input argument.
  • Boundary markers: The skill does not implement delimiters or provide explicit instructions to the agent to ignore embedded commands within user prompts.
  • Capability inventory: The skill utilizes the infsh tool to perform network requests for video/audio generation and to execute media processing tasks.
  • Sanitization: There is no evidence of input validation or sanitization applied to prompt strings before they are processed by the tool.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 22, 2026, 03:21 PM