check-link-wallet

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • Data Privacy and Minimization: The instructions prioritize user privacy by limiting the exposure of sensitive information. The agent is directed to provide only summaries, such as the brand and last four digits for payment methods, and city/postcode for shipping addresses by default, unless the user explicitly requests more detail.
  • Credential Protection: The skill specifically instructs the agent not to request parameters that would return live payment credentials (such as card or shared_payment_token) during a status check. This design ensures that highly sensitive data is not unnecessarily retrieved or exposed during routine inquiries.
  • Indirect Prompt Injection Surface: As the skill processes data from external sources (the connected Link account), it possesses an indirect prompt injection surface. However, the instructions to summarize data rather than dumping raw JSON, along with the read-only nature of the tools, act as effective mitigations against potentially malicious content within that data.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 03:49 AM