The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's internal scripts (scan.py, exploit.py, report.py) use subprocess.run to execute external security tools (Slither, Aderyn), build environments (Forge), and document generators (Pandoc). These calls are performed using list-based arguments without shell=True, which prevents command injection vulnerabilities.
[PROMPT_INJECTION]: As a code analysis tool, the skill processes untrusted Solidity source code. This creates a surface for indirect prompt injection where malicious comments or code patterns could attempt to influence the agent's manual semantic analysis. This is an inherent risk for auditing tools and is mitigated by the structured use of multiple automated scanners and explicit methodology.
Ingestion points: Solidity contract files read during the scope identification phase (Step 2).
Boundary markers: None used when presenting code for manual agent review.
Capability inventory: The skill has access to Bash and can execute generated tests via subprocess.run (Forge), providing a significant capability tier.
Sanitization: Automated tools provide objective findings before the agent performs subjective analysis.

solidity-audit