# Qwen Code S-Edition

AI-powered security and architecture scanner CLI for any codebase.

## Installation

```bash
npm install -g qwen-code-s-edition
```

## Authentication

```bash
qwen-s auth <your-api-key>
```
Get your API key at: https://supreme.silenceai.net/#pricing

## Usage

```bash
qwen-s [path] [options]
```

### Options

| Flag | Default | Description |
|------|---------|-------------|
| `--no-verify` | off | Skip LLM verification of findings (faster, more false positives) |
| `--verify-all` | off | Verify all findings regardless of count (uses more API quota) |
| `--min-severity <level>` | LOW | Minimum severity to report: CRITICAL \| HIGH \| MEDIUM \| LOW |
| `--no-llm` | off | Skip LLM architecture analysis (graph only, no AI recommendations) |
| `--skip-ua` | off | Use cached knowledge-graph.json (skip Understand-Anything rebuild) |
| `--no-arch` | off | Skip architecture scan entirely |
| `--no-security` | off | Skip security scan entirely |
| `--open` | off | Open HTML report in browser after scan |
| `--no-unicode` | off | Use ASCII-only output (for Windows CMD/PowerShell) |

### Dashboard subcommand

```bash
qwen-s dashboard --path /path/to/project
```

## Scan Phases

1. **Phase 1/4** — Security Scan (supreme-max + Qwen LLM verification)
2. **Phase 2/4** — Architecture Scan (tree-sitter or Understand-Anything)
3. **Phase 3/4** — Dockerfile Scan (hadolint — skipped if not installed)
4. **Phase 4/4** — Report Generation

Reports are uploaded to your dashboard at supreme.silenceai.net automatically.

## Requirements

- Node.js 18+
- API key from supreme.silenceai.net
- (Optional) hadolint for Dockerfile scanning