Output Format

argus generates a CLAUDE.md file with structured sections that AI assistants can easily parse and understand.

File Structure

The output file has two main regions:

<!-- ARGUS:AUTO -->
... auto-generated content ...
<!-- /ARGUS:AUTO -->

<!-- ARGUS:CUSTOM -->
... your custom notes (preserved on re-scan) ...
<!-- /ARGUS:CUSTOM -->

Auto-Generated Sections

Project Overview

Basic project information extracted from package.json, go.mod, pyproject.toml, etc.

Quick Reference

Prioritized commands for common tasks:

## Quick Reference

```bash
# Build
npm run build

# Test
npm test

# Lint
npm run lint

# Run
npm start

### Architecture

Project structure analysis:

```markdown
## Architecture

**Style:** Monorepo with Turborepo

**Entry Point:** `apps/web/src/main.tsx`

┌─────────────┐ │ apps │ └──────┬──────┘ │ ┌──────┴──────┐ │ packages │ └─────────────┘

Tech Stack

Languages, frameworks, and tools:

## Tech Stack

### Languages
- **TypeScript** 89.5%
- **CSS** 10.5%

### Frameworks
- React 18.2.0
- Next.js 14.0.0

### Tools
- ESLint
- Prettier
- GitHub Actions

Detected Patterns

Code patterns found in the codebase:

## Detected Patterns

### Data Fetching
- **React Query** - TanStack Query hooks (12 files)
- **fetch** - Native fetch API (8 files)

### State Management
- **useState** - React state (45 files)
- **Zustand** - Global state store (3 files)

Custom Section

Add your own notes that persist across scans:

<!-- ARGUS:CUSTOM -->
## My Notes

- Remember to run `docker-compose up` before testing
- API keys are in 1Password vault "Dev"
- Contact @john for database access
<!-- /ARGUS:CUSTOM -->

Output Indicators

argus uses these indicators in the output:

Indicator	Meaning
✅	Success
⚠️	Warning
🔍	Scanning/analyzing
→	Progress
🔄	Processing
📊	Analysis results
📄	File output
👁️	Watch mode
❌	Error