return-funds
Pass
Audited by Gen Agent Trust Hub on Mar 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access detected.
- [COMMAND_EXECUTION]: The skill uses specific, purpose-built wallet tools (e.g.,
wallet_send,payment_request_refund) for transactions rather than executing arbitrary or risky shell commands. - [PROMPT_INJECTION]: The instructions prioritize safety and explicit user confirmation, which is the opposite of a prompt injection attack. No override or bypass attempts are present.
- [DATA_EXFILTRATION]: The skill handles transaction data locally within the agent's wallet environment and does not attempt to exfiltrate sensitive data or credentials to external servers.
Audit Metadata