full-code-reviewlisted

# Full Code Review — Multi-Agent Deep Review Runs a comprehensive code review using parallel specialized agents. Each codebase gets 3 review dimensions (Security & Auth, Code Quality & Patterns, Business Logic). Results are synthesized into numbered markdown files in a `claude-code-review/` output folder. ## Quick Start ``` /full-code-review ``` Then specify which codebases to review (paths relative to workspace root). The agent will discover them automatically if the workspace has obvious top-level directories. ## Output Structure All output goes to `<workspace-root>/claude-code-review/`: ``` claude-code-review/ 00-executive-summary.md ← Severity counts, top priorities, phased fix plan 01-<codebase>-security.md ← Auth, injection, data exposure, secrets 02-<codebase>-code-quality.md ← Patterns, TypeScript, duplication, testing 03-<codebase>-business-logic.md ← Order flows, payments, scheduling, edge cases ... (repeats per codebase) ``` ## Review Dimensions Per Codebase ### Dimension 1: Security & Auth - Authentication vulnerabilities (JWT handling, token storage, session management) - Authorization flaws (missing guards, privilege escalation, role checks) - Injection attacks (SQL injection, command injection, XSS) - Data exposure (sensitive data in responses, PII leaks, console logging) - Input validation (missing validation, DTO gaps, unchecked inputs) - Rate limiting (missing limits on sensitive endpoints) - Payment security (webhook signature ver