blockchain-web3

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches security tools from industry-standard repositories.
  • Downloads Slither and Mythril from PyPI.
  • Fetches Foundry installation scripts from the official Paradigm domain.
  • References various blockchain explorer and decompiler services like Etherscan and Dedaub.
  • [REMOTE_CODE_EXECUTION]: Contains a command pattern for piping a remote script to the shell.
  • Evidence found in payloads.md and guides: curl -L https://foundry.paradigm.xyz | bash. This is a standard installation method for a well-known Ethereum development toolkit.
  • [COMMAND_EXECUTION]: Instructs the agent to execute various CLI tools for auditing.
  • Uses slither, myth, forge, cast, and anvil for static analysis and dynamic testing.
  • Executes compiler management tools like solc-select.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external smart contract code.
  • Ingestion points: Reads local Solidity and Vyper files during audits.
  • Boundary markers: The skill does not explicitly warn the agent to ignore instructions embedded in the audited code.
  • Capability inventory: Includes shell access (Bash), file write access, and network fetch.
  • Sanitization: Not explicitly present in the provided instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 04:36 PM