What problem does it solve? Code reviews often miss security regressions hidden in diffs, especially when reviewers skip git history or underestimate small changes. This Skill systematically analyzes PRs, commits, and diffs to detect removed validations, access control bypasses, reentrancy risks, and other vulnerabilities before they reach production. ## Core Features & Use Cases - Risk-Adaptive Analysis: Classifies changes by risk level (HIGH/MEDIUM/LOW) and adapts review depth to codebase size (SMALL/MEDIUM/LARGE) so effort goes where it matters. - Git History & Blast Radius: Uses git blame and log searches to detect security regressions, and quantifies blast radius by counting callers of modified functions. - Adversarial Modeling & Reporting: Builds concrete attacker models and exploit scenarios for HIGH RISK changes, then generates a structured markdown report with severity tables, findings, and recommendations. - Use Case: A smart contract PR removes a require(amount > 0) check. The Skill traces the check's origin via git blame, finds it was added to fix a prior exploit, rates exploitability as EASY, and flags the change as CRITICAL in the final report. ## Quick Start Ask the AI to run a differential security review on the current PR or commit range and generate the full markdown report.
