web3-audit
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous shell command examples using
grepandforge(Foundry) to assist in codebase analysis and exploit verification. These instructions are diagnostic in nature and intended for standard security research workflows. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and analysis of external smart contract source code (Solidity/Rust). While this creates an indirect prompt injection surface, the skill provides specific structured methodologies to guide the agent's analysis, which is fundamental to its auditing purpose.
- Ingestion points: Analysis of external smart contract files and function logic in Solidity and Rust.
- Boundary markers: Employs comparison tests between function families and logical boundary condition checks to structure the analysis.
- Capability inventory: Utilizes command-line search tools (
grep), contract testing frameworks (Foundry/forge), and static/symbolic analysis tools (Slither,Mythril). - Sanitization: Focused on vulnerability discovery and reporting; no untrusted code is executed in a production context.
Audit Metadata