beneficiary-creation

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes the airwallex CLI binary and utilizes an MCP server to perform banking operations.
  • It specifically uses airwallex auth login for session management and various API-linked commands for beneficiary management.
  • [EXTERNAL_DOWNLOADS]: The skill fetches real-time banking requirements and schemas from the vendor's official domain.
  • It utilizes WebFetch to access country-specific guidelines at https://www.airwallex.com/docs/payouts/payout-network/bank-accounts/{country-slug}.
  • [PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it is designed to parse and interpret data from untrusted external documents.
  • Ingestion points: Supplier invoices, contracts, and vendor lists uploaded by the user are processed to extract data fields (SKILL.md, Step 1).
  • Boundary markers: No specific delimiters or instructions are provided to the agent to treat document content as purely passive data.
  • Capability inventory: The skill can list, search, create, update, and validate beneficiary records via the Airwallex infrastructure.
  • Sanitization: The skill mitigates risks by requiring explicit human confirmation of extracted data and using mandatory validation API calls before any creation event (SKILL.md, Steps 10-12).
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 09:07 AM
Security Audit — agent-trust-hub — beneficiary-creation