What problem does it solve? Smart contract bugs can drain funds, yet manual security reviews are slow and expensive. This Skill runs an automated security audit of Solidity code while you develop, surfacing high-confidence findings in minutes before every commit. ## Core Features & Use Cases - Parallel Multi-Agent Scanning: Orchestrates four vector-scanning agents that each check the full codebase against a 170-item attack vector reference covering reentrancy, oracle manipulation, proxy storage collisions, ERC4626 inflation attacks, and more. - Deep Adversarial Mode: The deep mode spawns an additional adversarial reasoning agent that hunts for logic errors, economic exploits, and multi-transaction attack paths beyond pattern matching. - Confidence-Scored Reports: Every finding passes a false-positive gate, receives a confidence score, and is merged, deduplicated, and formatted into a structured markdown report with diff-style fixes. - Use Case: Before deploying a new vault contract, run the audit on src/Vault.sol to catch missing access controls, unchecked return values, and reentrancy issues, then write the findings to a report file with the --file-output flag. ## Quick Start Ask the agent to run a security audit on your Solidity repository, or pass a specific contract filename to review just that file.
