linkfox-seerfar-ozon-market-keyword-search

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes and displays marketplace keyword data (e.g., Russian search terms and product titles) retrieved from external marketplace APIs. This content is ingested into the agent's context, creating a surface where adversarial text within marketplace metadata could potentially influence the agent's behavior.
  • Ingestion points: API response data processed in scripts/seerfar_ozon_market_keyword_search.py.
  • Boundary markers: None implemented to distinguish external data from instructions.
  • Capability inventory: File system writes to the project directory and network operations to vendor APIs.
  • Sanitization: Relies on standard JSON parsing without specific sanitization for embedded natural language instructions.
  • [EXTERNAL_DOWNLOADS]: The skill performs network requests to vendor-owned domains to provide marketplace data, authentication, and billing services.
  • Domains: tool-gateway.linkfox.com, api.linkfox.com, and agent-api.linkfox.com.
  • Purpose: Market keyword search, user login/onboarding, and plan subscription management.
  • [COMMAND_EXECUTION]: The skill's documentation and scripts provide instructions for the user to execute shell commands for configuration and dependency management.
  • Evidence: references/onboarding.md provides commands for persistent environment variable configuration in shell profiles (.bashrc, .zshrc, setx). scripts/onboarding.py and scripts/seerfar_ozon_market_keyword_search.py mention installing dependencies like requests, qrcode, and pillow via pip.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 07:02 AM
Security Audit — agent-trust-hub — linkfox-seerfar-ozon-market-keyword-search