yeepay-payment-integration
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill dynamically fetches API documentation and JSON schemas from YeePay's official domain (
open.yeepay.com) to ensure it uses the latest field specifications. It also fetches the latest Java SDK version from Maven Central (central.sonatype.com). These are standard operational behaviors for a vendor-provided integration tool. - [COMMAND_EXECUTION]: Several Python scripts in the
scripts/directory use thesubprocessmodule to call the system'sopensslbinary. This is used for legitimate tasks such as parsing X.509 certificates and exporting public keys from PEM files. Additionally, validation scripts usesubprocessto run local test suites for verifying cryptographic implementation correctness. - [CREDENTIALS_UNSAFE]: The skill contains hardcoded private keys within its test vector directories (
scripts/rsa/tests/vectors/andscripts/sm/tests/vectors/). These keys are explicitly documented as being for verification purposes only and are prohibited from use in real environments. They are necessary for ensuring the local signature logic matches the platform's expectations. - [INDIRECT_PROMPT_INJECTION]: The skill has a potential indirect injection surface because it instructions the agent to fetch and process remote markdown documentation from
open.yeepay.com. This is an inherent risk of dynamic documentation fetching, though the source is the official vendor portal and the behavior is essential for the skill's primary function.
Audit Metadata