> ## Documentation Index
> Fetch the complete documentation index at: https://arclux-os.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# How to Use ARCLUX

> CLI, daemon, web dashboard, and VS Code extension

Everything ARCLUX can do today, from a one-off CLI check to always-on background analysis with editor integration.

## 1. One-off CLI commands

Point any command at a local repo path (defaults to .):

```bash theme={null}
arclux analyze [path]
arclux graph [path]
arclux impact <file> [path]
arclux doctor [path]
arclux diagnose [path]
arclux diff <from> <to> [path]
arclux verify [path]
arclux security [path]
arclux search <query> [path]
arclux script <file.arclux>
arclux config [path]
arclux shell
arclux mcp
```

diagnose output includes clickable file paths (OSC 8 hyperlinks) in terminals that support it (Termux, iTerm2, VS Code integrated terminal) -- tap/click to open the file directly.

<Note>
  `analyze` also accepts a remote repo URL (arclux analyze [https://github.com/org/repo](https://github.com/org/repo)) -- it clones, analyzes, and cleans up, with SSRF/private-network guards on the URL before anything else.
</Note>

## 2. DSL scripting (arclux script)

Scripts chain every engine capability in one readable file:

```arclux theme={null}
repo = analyze("~/flask")
impact(repo, "app.py")
check(repo, "orphanFiles")
```

Built-ins: analyze, doctor, check, graph, callgraph, impact, search, security, diff, archdiff + helpers (len, sum, filter, sort, exists, keys, values, env, cwd, extensions, checkids). extensions() and checkids() grow automatically as new parsers/detectors are registered -- the DSL is registry-driven, it never needs a code change to know a new language.

## 3. Always-on daemon

Instead of re-running commands by hand, start ARCLUX as a background process that watches your repo and re-analyzes on every save:

```bash theme={null}
arclux daemon --detach
arclux daemon --status
arclux daemon --stop
```

Run without --detach to see live output in the foreground instead.

The daemon exposes a local HTTP+SSE bridge so any editor or terminal can connect:

```bash theme={null}
curl http://127.0.0.1:PORT/analysis
curl http://127.0.0.1:PORT/events
```

Find the port from \~/.arclux/endpoints/DAEMON\_ID.json, written automatically when the daemon starts.

## 4. Web dashboard

```bash theme={null}
cd apps/web
pnpm run dev
```

Open localhost:3000/ORG/REPO for a given GitHub URL. From the Overview page:

* Click any file in the Project structure tree to open it
* The File tab shows syntax-highlighted source (Python, JavaScript, TypeScript today) with inline diagnostic gutter markers -- a colored bar on any line with a finding, click it to expand the message + fix suggestion
* Dependencies and Impact tabs show what a file needs and what breaks if you change it
* Click the collapse chevron to give the file panel full width

The Graph page renders the full dependency graph; selecting a node opens a focus panel showing what it needs and what it affects.

## 5. VS Code extension

Connects to a running daemon and surfaces diagnostics in VS Code's Problems panel + a status bar module count.

```bash theme={null}
cd apps/vscode-extension
pnpm install && pnpm build
```

Then load it via VS Code's Extension Development Host. Requires a daemon already running for the workspace folder.

## 5. Audit — the findings theater

Run the full analysis (doctor + security + attack surface) as one guided
experience:

* `/{org}/{repo}/audit` — standalone page, repo prefilled from the URL
* or the `▸ audit` tab inside `/script`

The terminal streams the scan (systemd-style boot, real numbers), then
chapters reveal one finding at a time in the FOCUS panel with a file
preview overlay. On the graph page, **run audit on graph** replays every
finding as a breathing severity halo on the 3D nodes — critical red,
high orange, medium yellow.