product-relighting

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides setup instructions to add a Model Context Protocol (MCP) server from the vendor's domain at https://mcp.apimage.org/mcp. This is a vendor-owned resource used to provide the skill's core functionality.
  • [COMMAND_EXECUTION]: A configuration command is provided for the user to register the necessary tool transport in their environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted image data through analysis and editing tools.
  • Ingestion points: Images ingested via analyze_image and edit_image in SKILL.md.
  • Boundary markers: The skill advises using explicit instructions like "keep the product, composition and framing unchanged" to constrain the model's behavior.
  • Capability inventory: Tools for image analysis, image editing, and brand asset creation.
  • Sanitization: The skill relies on natural language constraints to guide tool output when processing external content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 12:36 AM