ai-image-generation

Pass

Audited by Gen Agent Trust Hub on Apr 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the infsh CLI tool to perform AI image generation. Shell access is restricted to this specific command via the allowed-tools frontmatter configuration, preventing execution of unauthorized system commands.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves assets and links to documentation on inference.sh and its subdomains. These resources belong to the official service provider and are categorized as safe.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it passes user-defined prompts to external AI models through the CLI tool.
  • Ingestion points: User prompts are accepted as values in the JSON --input argument within SKILL.md.
  • Boundary markers: The skill utilizes JSON keys to encapsulate user input, providing a structural boundary.
  • Capability inventory: Interactions are limited to the infsh tool which communicates with a managed API.
  • Sanitization: Input is passed to external AI models that implement their own content safety and filtering mechanisms.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 25, 2026, 06:52 PM
Security Audit — agent-trust-hub — ai-image-generation