Audit Command

Purpose: Internal codebase audit for architecture understanding, technical debt assessment, and dependency analysis

Workflow

Step 1: Validate Target

• Check that target directory exists
• Determine if it's a full codebase or specific component
• Identify project type (Node.js, Python, PHP, Go, etc.)

Step 2: Launch Parallel Auditors

Start 5 specialized agents in parallel:

1. structure-mapper: Map directory structure and organization
2. dependency-analyzer: Analyze external dependencies
3. pattern-detector: Identify code patterns and conventions
4. tech-debt-assessor: Assess technical debt and code smells
5. documentation-generator: (waits for others, runs last)

Each agent operates independently and returns their findings.

Step 3: Consolidate Findings

Merge outputs from all agents:

• Structure map from structure-mapper
• Dependency analysis from dependency-analyzer
• Pattern inventory from pattern-detector
• Tech debt report from tech-debt-assessor

Step 4: Generate Documentation

Invoke documentation-generator with consolidated findings to create:

• ARCHITECTURE.md - System architecture overview
• COMPONENTS.md - Component documentation
• DEPENDENCIES.md - Dependency analysis
• TECHNICAL_DEBT.md - Tech debt assessment
• AUDIT_REPORT.md - Complete audit summary

Step 5: Present Findings

Create final summary report highlighting:

• Key architectural insights
• Critical technical debt items
• Recommended next steps
• Onboarding guide for new developers

Output Format

The audit produces:

audit-output/
├── ARCHITECTURE.md
├── COMPONENTS.md
├── DEPENDENCIES.md
├── TECHNICAL_DEBT.md
└── AUDIT_REPORT.md

Each file contains structured, actionable information about YOUR codebase.

Usage

/audit "target/directory"

Or for full codebase:

/audit

Notes

• Audit is read-only (no code changes)
• Analyzes INTERNAL code (your existing codebase)
• Focus is on understanding what you have, not evaluating external options
• All outputs are in markdown format
• Findings are fact-based, not opinionated