x9-wb-product-search
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
agent-edgecommand-line tool (specificallyagent-edge --session main tab new) as a fallback mechanism to manage browser tabs within the user's Edge session. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of product descriptions and user-generated reviews from Wildberries, which creates a potential surface for indirect instructions to be passed to the agent.
- Ingestion points: The agent is instructed to read product detail pages and extensive customer review sections on Wildberries to perform audits (referenced in
SKILL.md§6 andreferences/review-audit.md). - Boundary markers: Absent; there are no specific instructions or markers provided to the agent to treat external review content as potentially malicious or to ignore instructions embedded within those reviews.
- Capability inventory: The skill has the ability to interact with the browser via CDP, add items to the user's favorites, and modify the shopping cart (described in the Tools section of
SKILL.md). - Sanitization: Absent; the skill does not specify any methods for sanitizing, escaping, or filtering instructions from the ingested review data before it is processed by the agent.
Audit Metadata