Add top-level CLI tab and intro page feature section

cli/authentication.mdx

@@ -0,0 +1,98 @@
+---
+title: "Authentication"
+description: "How the Linkrunner CLI authenticates with your account"
+icon: "key"
+---
+
+The CLI uses **API keys** to authenticate with the Linkrunner API. There are two ways to obtain a key: the device login flow (recommended) or manual key creation from the dashboard.
+
+## Device Login Flow
+
+When you run `lr login`, the CLI uses an OAuth2 Device Authorization flow:
+
+1. The CLI requests a **user code** from the Linkrunner API
+2. You are shown a code (e.g. `BVKD-3NH7`) and a URL to open in your browser
+3. You open the URL, log into the Linkrunner dashboard, and authorize the code
+4. The CLI polls for approval and receives an API key automatically
+
+```
+$ lr login
+
+  Your device code: BVKD-3NH7
+
+  Open this URL to authorize:
+  https://dashboard.linkrunner.io/cli/auth?code=BVKD-3NH7
+
+  Waiting for authorization...
+
+  ✔ Logged in successfully!
+```
+
+<Note>
+The device code expires after **10 minutes**. If it expires, run `lr login` again to get a new code.
+</Note>
+
+## API Keys
+
+Each login creates an API key that is stored locally. Keys have the prefix `lr_cli_` and can be managed from the Linkrunner dashboard.
+
+### Key Properties
+
+| Property | Description |
+|---|---|
+| **Prefix** | Starts with `lr_cli_` (first 12 chars visible in dashboard) |
+| **Label** | Optional name for identification (auto-set to "CLI Login" for device flow) |
+| **Last Used** | Timestamp of the most recent API call |
+| **Created** | When the key was generated |
+
+### Managing Keys from the Dashboard
+
+Go to **Settings > CLI Keys** in the Linkrunner dashboard to:
+
+- **View** all active CLI API keys and when they were last used
+- **Create** a new key manually (useful for CI/CD or shared environments)
+- **Revoke** a key to immediately invalidate it
+
+<Warning>
+Revoking a key is permanent. Any CLI session using that key will need to re-authenticate with `lr login`.
+</Warning>
+
+### Creating Keys for CI/CD
+
+For automated environments where interactive login isn't possible, create a key manually:
+
+1. Go to **Settings > CLI Keys** in the dashboard
+2. Click **Create New Key**
+3. Copy the key (it's only shown once)
+4. Set it as an environment variable:
+
+```bash
+export LINKRUNNER_API_KEY=lr_cli_...
+```
+
+The CLI automatically picks up the `LINKRUNNER_API_KEY` environment variable when no stored session exists.
+
+## Credential Storage
+
+Your API key is stored in a local config file managed by [conf](https://github.com/sindresorhus/conf):
+
+- **macOS:** `~/Library/Preferences/linkrunner-nodejs/config.json`
+- **Linux:** `~/.config/linkrunner-nodejs/config.json`
+- **Windows:** `%APPDATA%/linkrunner-nodejs/config.json`
+
+## Logout
+
+To clear your stored credentials:
+
+```bash
+lr logout
+```
+
+This removes the locally stored API key but does **not** revoke it on the server. To fully revoke access, also delete the key from **Settings > CLI Keys** in the dashboard.
+
+## Security
+
+- API keys are hashed (SHA-256) before storage on the server — the raw key is never stored
+- Keys are scoped to your user account, not individual projects
+- The device flow codes use an ambiguity-excluded character set to avoid typos
+- Device codes are single-use and expire after 10 minutes

cli/ci-cd.mdx

@@ -0,0 +1,168 @@
+---
+title: "CI/CD Integration"
+description: "Use the Linkrunner CLI in your CI/CD pipelines"
+icon: "rotate"
+---
+
+## Overview
+
+The Linkrunner CLI supports CI/CD workflows through two key features:
+
+- **`--ci` flag** -- commands exit with code 0 (pass) or 1 (fail) instead of remaining interactive
+- **`--json` flag** -- structured JSON output for machine parsing
+
+These flags are available on `lr doctor`, `lr validate`, `lr analyze`, and `lr test`.
+
+## Exit Codes
+
+| Code | Meaning |
+|---|---|
+| `0` | All checks passed (no errors) |
+| `1` | One or more errors found |
+
+By default, warnings do not cause a non-zero exit. Add `--fail-on-warn` to treat warnings as failures.
+
+## Commands for CI
+
+### lr doctor --ci
+
+Validates platform configuration and source code:
+
+```bash
+# Fail on errors only
+lr doctor --ci
+
+# Fail on errors AND warnings
+lr doctor --ci --fail-on-warn
+```
+
+### lr test --ci
+
+Verifies API connectivity and token validity:
+
+```bash
+lr test --ci
+```
+
+### lr analyze --ci
+
+Runs standard checks plus AI-powered deep analysis:
+
+```bash
+lr analyze --ci --fail-on-warn
+```
+
+## Pipeline Examples
+
+### GitHub Actions
+
+```yaml
+name: Linkrunner SDK Check
+on: [push, pull_request]
+
+jobs:
+  sdk-check:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-node@v4
+        with:
+          node-version: 20
+
+      - name: Install Linkrunner CLI
+        run: npm install -g linkrunner-cli
+
+      - name: Validate SDK integration
+        run: lr doctor --ci --fail-on-warn
+
+      - name: Test API connectivity
+        run: lr test --ci
+```
+
+### GitLab CI
+
+```yaml
+sdk-check:
+  image: node:20
+  script:
+    - npm install -g linkrunner-cli
+    - lr doctor --ci --fail-on-warn
+    - lr test --ci
+```
+
+### Bitbucket Pipelines
+
+```yaml
+pipelines:
+  default:
+    - step:
+        name: SDK Validation
+        image: node:20
+        script:
+          - npm install -g linkrunner-cli
+          - lr doctor --ci --fail-on-warn
+```
+
+## JSON Output
+
+Use `--json` to get machine-readable output for custom processing:
+
+```bash
+lr doctor --json | jq '.summary.errors'
+```
+
+The JSON structure for `lr doctor --json`:
+
+```json
+{
+  "project": {
+    "type": "flutter",
+    "sdkVersion": "3.2.1",
+    "root": "/path/to/project"
+  },
+  "results": [
+    {
+      "id": "check-id",
+      "name": "Check name",
+      "status": "pass",
+      "severity": "error",
+      "message": "Description",
+      "fix": "Suggested fix",
+      "autoFixable": false,
+      "docsUrl": "https://..."
+    }
+  ],
+  "summary": {
+    "passed": 5,
+    "warnings": 1,
+    "errors": 0
+  }
+}
+```
+
+## Combining Flags
+
+You can combine `--ci` and `--json` for both structured output and exit codes:
+
+```bash
+# Get JSON output AND exit code
+lr doctor --ci --json > results.json
+echo "Exit code: $?"
+```
+
+## Requirements
+
+- The project must contain a `.linkrunner.json` file (for `lr test`) or recognizable project files (for `lr doctor`)
+- No interactive prompts are shown when `--ci` or `--json` is used
+- `lr test` and `lr status` require a `.linkrunner.json` file created by `lr init`
+
+<Note>
+For CI environments, run `lr init` locally first and commit the `.linkrunner.json` file to your repository. The CLI reads this file to identify your project during CI runs.
+</Note>
+
+## Related
+
+- [lr doctor](/cli/commands/doctor) -- full diagnostics reference
+- [lr test](/cli/commands/test) -- API connectivity checks
+- [lr analyze](/cli/commands/analyze) -- AI-powered analysis

cli/commands/analyze.mdx

@@ -0,0 +1,105 @@
+---
+title: "lr analyze"
+description: "Run AI-powered deep analysis of your SDK integration"
+icon: "robot"
+---
+
+## Usage
+
+```bash
+lr analyze [options]
+```
+
+## Description
+
+Runs all standard `lr doctor` checks plus an AI-powered deep analysis of your codebase. This is equivalent to running `lr doctor --deep`.
+
+The AI analysis inspects your source code for subtle integration issues, incorrect usage patterns, missing best practices, and platform-specific problems that pattern-based checks cannot catch.
+
+<Note>Requires authentication. Run `lr login` first.</Note>
+
+## Options
+
+| Flag | Description |
+|---|---|
+| `--json` | Output results as JSON |
+| `--ci` | Exit with code 0 (pass) or 1 (fail) |
+| `--fail-on-warn` | Treat warnings as failures in CI mode |
+
+## What It Does
+
+1. Runs all standard platform configuration checks (same as `lr doctor`)
+2. Runs source code pattern scanning (same as `lr doctor`)
+3. Sends project context to an AI model for deep analysis
+4. Reports any additional issues found by the AI, including:
+   - File path and line number of the issue
+   - Severity level (error, warn, or info)
+   - Suggested fix
+
+## Examples
+
+### Basic analysis
+
+```bash
+lr analyze
+```
+
+```
+Linkrunner Doctor
+Diagnosing your SDK integration...
+
+  Project: react-native (SDK v2.3.0)
+  Root: /path/to/project
+
+── Platform Configuration ──
+  ✔ rn-linkrunner found in package.json
+  ...
+
+── Source Code ──
+  ✔ init() call found
+  ...
+
+── Deep Analysis (AI-powered) ──
+  ✘ [src/App.tsx:15] init() called outside useEffect - may run on every render
+    Fix: Move init() inside a useEffect with an empty dependency array
+  ⚠ [src/screens/Payment.tsx:42] capturePayment amount is hardcoded
+    Fix: Use the actual transaction amount from your payment provider
+  ✔ No additional issues found by AI analysis.
+```
+
+### JSON output for scripting
+
+```bash
+lr analyze --json
+```
+
+The JSON output includes a `deepAnalysis` array in addition to the standard `results`:
+
+```json
+{
+  "project": { "type": "react-native", "sdkVersion": "2.3.0", "root": "..." },
+  "results": [...],
+  "summary": { "passed": 5, "warnings": 1, "errors": 1 },
+  "deepAnalysis": [
+    {
+      "file": "src/App.tsx",
+      "line": 15,
+      "severity": "error",
+      "message": "init() called outside useEffect",
+      "fix": "Move init() inside a useEffect with an empty dependency array"
+    }
+  ]
+}
+```
+
+### In CI
+
+```bash
+lr analyze --ci --fail-on-warn
+```
+
+## Related
+
+- [lr doctor](/cli/commands/doctor) -- standard diagnostics without AI analysis
+- [lr suggest](/cli/commands/suggest) -- AI-powered feature suggestions
+- [CI/CD Integration](/cli/ci-cd) -- pipeline examples