What problem does it solve? Manual smart contract security review is slow and expensive, while most automated scanners only flag bad lines and miss accounting bugs where a tracked total drifts from the value it claims to represent. This Skill builds the protocol's money model first, then attacks it systematically to find who is left holding the loss. ## Core Features & Use Cases - Accounting-first analysis: Builds a money map of assets, tracked totals, invariants, lifecycles, and actor cohorts before hunting for bugs, targeting the desync class behind roughly a third of real High findings. - Twelve parallel attack lenses: Runs independent subagents covering accounting-desync, share-exchange-rate, temporal-cohort, liquidation-solvency, cross-chain-state, rounding, MEV, DoS, access control, integrations, edge states, and flow completeness. - Dedup, judging, and calibrated reporting: Merges duplicate findings with hard gates for function isolation and mechanism preservation, applies four severity gates, and emits a report with Description and Recommended Mitigation per finding. - Use Case: Point it at a DeFi vault repository before a paid audit to catch missing decrements, one-sided branch updates, and last-withdrawer insolvency bugs early. ## Quick Start Ask your agent to run 0xSimao AI on this repository to perform a full accounting-first security audit of the in-scope Solidity contracts.