skill-03-defi-operations
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns were detected. The skill follows industry security standards for blockchain interactions, such as using scoped approvals and screening counterparties.\n- [PROMPT_INJECTION]: The skill interacts with external blockchain data and APIs, creating an inherent attack surface for indirect prompt injection.\n
- Ingestion points: External quotes and protocol data from
defi-protocol-server, 1inch API, and protocol subgraphs.\n - Boundary markers: The skill does not explicitly define isolation markers for the ingested data.\n
- Capability inventory: Transaction simulation and smart contract execution via
blockchain-rpc-serverandskill-02.\n - Sanitization: Employs comprehensive mitigation strategies, including address allowlisting, price impact validation, and mandatory transaction simulation prior to execution.
Audit Metadata