evm-swiss-knife
Fail
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPERSISTENCE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute the Foundry installation script via
curl -L https://foundry.paradigm.xyz | bash. This represents the official installation method for the required toolset. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external blockchain data and RPC responses, which may contain untrusted content.
- Ingestion points: Data retrieved from RPC endpoints and blockchain state via
cast calloperations inSKILL.md. - Boundary markers: None identified; output is processed directly by the agent.
- Capability inventory: Shell command execution (
cast), file system interaction for ABIs, and network requests to user-defined RPC URLs. - Sanitization: No sanitization or validation of the retrieved blockchain data is implemented.
- [DATA_EXFILTRATION]: The skill facilitates the use of sensitive private keys and interacts with external RPC providers. While it advises using environment variables, the use of sensitive keys in CLI parameters poses a risk of exposure to untrusted endpoints.
- [COMMAND_EXECUTION]: The skill uses the
castcommand-line tool to perform blockchain queries and transactions. - [PERSISTENCE]: The installation process involves instructions for modifying and sourcing shell profile files such as
~/.bashrcand~/.zshrc.
Recommendations
- HIGH: Downloads and executes remote code from: https://foundry.paradigm.xyz - DO NOT USE without thorough review
Audit Metadata