image

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides standard command-line examples for image optimization and analysis using tools like cwebp, mogrify (ImageMagick), jpegoptim, and curl. These are documented as quick optimization commands for the user's workflow.
  • [EXTERNAL_DOWNLOADS]: References official documentation, API guides, and pricing pages for well-known technology providers including Google (Gemini), OpenAI (GPT Image), Black Forest Labs (Flux), and Ideogram. These are trusted services for image generation.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest local context from files like .agents/product-marketing.md to inform its visual content production. This is a standard design pattern for maintaining brand consistency through local context ingestion.
  • [METADATA_POISONING]: Metadata fields (name, description, version) are consistent with the skill's stated purpose of producing professional visual assets.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 02:38 PM