ai-ecommerce-expert-koc-seeding-collage

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilizes npx to dynamically fetch and execute the @infimind/ecom-content-cli package. This is an official package associated with the vendor (Infimind) and is used for its intended e-commerce content generation purpose.
  • [COMMAND_EXECUTION]: The helper script scripts/imiva_mcp.py uses subprocess.Popen to bridge communication between the agent and the Node.js-based MCP client. The command is restricted to the specific npm package and does not use a shell (shell=False is implied by list arguments), which reduces the risk of command injection.
  • [CREDENTIALS_SAFE]: The instructions and configuration examples correctly emphasize that sensitive API tokens (MCP_TOKEN) should be stored in environment variables or a local configuration file, not in the skill's source code or shared logs.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: Product names, categories, and selling points are ingested via the create_koc_collage_task tool.
  • Boundary markers: The skill uses structured JSON objects to pass arguments to the MCP bridge script, providing clear data boundaries.
  • Capability inventory: The skill has the ability to execute the specific Infimind MCP CLI and perform file system reads for material paths.
  • Sanitization: Inputs are serialized into JSON in imiva_mcp.py before being passed to the subprocess, preventing simple injection attacks against the command-line interface.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 08:01 AM
Security Audit — agent-trust-hub — ai-ecommerce-expert-koc-seeding-collage