x9-wb-product-search

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Wildberries (product descriptions and reviews), which creates an inherent attack surface for indirect prompt injection.
  • Ingestion points: Wildberries search results, product pages, and customer reviews are read into the agent's context (SKILL.md §4, §6).
  • Boundary markers: The skill does not explicitly define delimiters to isolate untrusted web content from its instructions.
  • Capability inventory: The skill uses browser navigation via Chromium/Edge and has the ability to spawn subagents for verification (SKILL.md §8).
  • Sanitization: No explicit filtering or sanitization of instructions within the fetched data is mentioned.
  • [DYNAMIC_EXECUTION]: The skill utilizes a verification workflow that delegates analysis to a "fresh subagent" (Claude Opus or native Codex subagent) to ensure objectivity. This is a standard agentic pattern and is used here for safety verification (SKILL.md §8).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 07:41 AM
Security Audit — agent-trust-hub — x9-wb-product-search