ecommerce-ozon-brand-products

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a local Python script (scripts/ozon_brand_products.py) to facilitate communication with the NexScope API. This script is invoked with JSON parameters to perform data retrieval and processing for marketplace research.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external marketplace data from the Ozon API, creating a surface for potential indirect prompt injection if the source data were to contain malicious instructions.
  • Ingestion points: External market data is fetched via HTTP POST requests and stored in JSON format within the project's subdirectories (SKILL.md, scripts/ozon_brand_products.py).
  • Boundary markers: The skill does not implement specific delimiters or 'ignore' instructions for the fetched data, relying instead on the agent's ability to parse structured JSON.
  • Capability inventory: The skill has permissions to write files to the project directory for caching and data storage, and network access to the vendor's API proxy.
  • Sanitization: Employs standard JSON serialization and suggests using targeted extraction tools like jq to manage the amount of external content loaded into the LLM context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:14 AM
Security Audit — agent-trust-hub — ecommerce-ozon-brand-products