skill-02-smart-contract-interaction
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill provides technical instructions for blockchain interaction using standard tools and practices, emphasizing security verification steps.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface as it processes external data from blockchain RPC providers and block explorers.
- Ingestion points: Reads contract state via
eth_call, ABI definitions from external resources like Etherscan, and bytecode from various EVM chains. - Boundary markers: The instructions explicitly require contract verification and screening via a
security-auditMCP before any write interaction. - Capability inventory: The skill allows the agent to execute state-changing transactions, estimate gas, and deploy new contracts on multiple networks.
- Sanitization: Employs defensive measures including transaction simulation, bytecode hash verification, rug-pull pattern detection, and sanctions list checks to sanitize inputs from untrusted external contracts.
Audit Metadata