dripping-faucet
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows security best practices by explicitly instructing the agent never to log private keys and to manage secrets using environment variables or secure keystores.
- [PROMPT_INJECTION]: The skill includes specific defensive logic to mitigate indirect prompt injection by defining a strict trust boundary; it commands the agent to treat API response fields as data only and never as instructions to be followed.
- [COMMAND_EXECUTION]: Uses standard blockchain CLI tools (Foundry/cast) and provides clear guidance on how to use them without exposing sensitive information in process listings or shell history.
- [DATA_EXFILTRATION]: All network requests are directed to official Radius Network endpoints on the vendor-controlled domain (radiustech.xyz) for faucet and RPC operations.
Audit Metadata