skills/alipay/ai/alipay-payment-skill/Gen Agent Trust Hub

alipay-payment-skill

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external HTTP 402 responses, specifically the Payment-Needed header and amtPaymentLink parameter. This content is used as input for shell commands executed via the alipay-bot tool.
  • Ingestion points: references/402-payment.md (extracts headers and links from 402 responses); SKILL.md (extracts order numbers and transaction IDs from tool outputs).
  • Boundary markers: Instructions mandate wrapping shell arguments in single quotes and saving sensitive header data to isolated files before use.
  • Capability inventory: Execution of shell commands (alipay-bot), network proxying (curl-proxy), and package installation (npm/npx).
  • Sanitization: The skill includes logic to escape single quotes within variable values before they are interpolated into shell commands to prevent command injection.
  • [COMMAND_EXECUTION]: The skill relies on executing various subcommands of the alipay-bot CLI to perform business logic, including 402-buyer-pay, submit-payment, query-payment-status, and problem-feedback.
  • [EXTERNAL_DOWNLOADS]: A recovery mechanism in references/command-recovery.md allows for the installation of the @alipay/agent-payment package from the official npm registry. This process is protected by a mandatory SHA-512 integrity check (dist.integrity) to ensure the downloaded package matches the expected vendor-signed version.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 01:18 PM