mpstats

Warn

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Several shell scripts in the scripts/ directory unsafely interpolate shell variables directly into Python or Node.js code strings for URL encoding purposes. This pattern allows for arbitrary command execution if an attacker can influence the arguments passed to the script (e.g., brand names or category paths).
  • Evidence: Scripts such as scripts/ozon/ozon-brand.sh, scripts/wb/wb-brand.sh, scripts/wb/wb-category.sh, and scripts/ym/ym-seller.sh use the pattern python3 -c "... '$VAR' ..." instead of using safe argument passing like sys.argv.
  • [PROMPT_INJECTION]: The skill processes untrusted data fetched from marketplace APIs, including product reviews and comments, which constitutes an indirect prompt injection attack surface.
  • Ingestion points: scripts/wb/wb-sku.sh and scripts/ozon/ozon-sku.sh (fetching product comments).
  • Boundary markers: Absent.
  • Capability inventory: Shell command execution (bash), Network operations (curl to mpstats.io), and Local filesystem writes (config/.env).
  • Sanitization: Absent; the scripts return raw API data without filtering for embedded instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 7, 2026, 08:11 AM