evm-proxy-upgrade-analyzer

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [SAFE]: The skill operates as intended for its primary purpose of blockchain analysis. It does not attempt to access sensitive local files, credentials, or maintain persistence on the system.
  • [COMMAND_EXECUTION]: The script scripts/main.py provides a command-line interface that accepts user parameters through flags or standard input. It includes input sanitization via the _normalize_address function which uses a strict regular expression to validate hexadecimal Ethereum addresses.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the requests library to perform network operations to an external RPC endpoint provided by the user. This network access is necessary for the tool to fetch contract code and storage state from the blockchain.
  • [DATA_EXFILTRATION]: While the tool ingests external data from the specified RPC URL, this is the core functionality of the skill. The data processed (contract bytecode and storage words) is treated as technical values and is not interpolated into natural language instructions for the agent, mitigating risks associated with indirect prompt injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 07:53 PM
Security Audit — agent-trust-hub — evm-proxy-upgrade-analyzer