agent-wallet-policy
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses official vendor resources (circle.com domains and circlefin GitHub repositories) for documentation and references.
- [CREDENTIALS_UNSAFE]: The skill implements secure credential handling by explicitly instructing the agent not to handle or store OTPs. It ensures a human-in-the-loop process where the user provides secrets directly to the CLI prompt.
- [COMMAND_EXECUTION]: The skill utilizes the vendor's circle CLI tool for intended wallet management tasks. It distinguishes between safe read-only operations performed by the agent and interactive operations delegated to the user.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided spending limit values but mitigates injection risks by requiring the agent to confirm these values with the user and by having the user manually execute the final commands.
Audit Metadata