filtmall-shopping

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a bundled CommonJS file (assets/filtalgo-cli.cjs) executed via a Node.js wrapper (scripts/filtalgo.js). This local execution mechanism is the primary interface for the agent to perform marketplace tasks such as searching for products, managing carts, and querying logistics.
  • [INDIRECT_PROMPT_INJECTION]: The skill handles data from external shopping APIs which could potentially contain untrusted content. To mitigate this risk, it implements an extensive set of safety rules within the references/ directory. These rules include mandatory medical safety checks (stopping the recommendation process if allergic symptoms are described) and explicit boundary markers that force the agent to verify data integrity and obtain user confirmation before executing any modification to account or transaction states.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 03:47 PM
Security Audit — agent-trust-hub — filtmall-shopping