@agentuity/cli 3.0.0-alpha.7 → 3.0.0-beta.0

package/src/keychain.ts

@@ -1,205 +0,0 @@
-/**
- * macOS Keychain integration for secure auth token storage
- *
- * Stores auth tokens encrypted in the macOS Keychain using a per-device AES-256 key.
- * No user prompts required - fully automatic and secure.
- */
-
-const SERVICE_PREFIX = 'com.agentuity.cli';
-const KEY_ACCOUNT = 'aes-encryption-key';
-const AUTH_ACCOUNT = 'auth-token';
-const CODER_API_KEY_ACCOUNT = 'coder-hub-api-key';
-
-/**
- * Check if we're running on macOS
- */
-export function isMacOS(): boolean {
-	return process.platform === 'darwin';
-}
-
-/**
- * Get or create an AES encryption key stored in the macOS Keychain
- */
-async function ensureEncryptionKey(service: string): Promise<Uint8Array> {
-	// Try to read existing key
-	const find = Bun.spawn(
-		['security', 'find-generic-password', '-s', service, '-a', KEY_ACCOUNT, '-w'],
-		{ stderr: 'ignore' }
-	);
-
-	const stdout = await new Response(find.stdout).text();
-
-	if (stdout.length > 0) {
-		const b64 = stdout.trim();
-		return Uint8Array.from(Buffer.from(b64, 'base64'));
-	}
-
-	// Create a new 32-byte (256-bit) AES key
-	const key = crypto.getRandomValues(new Uint8Array(32));
-	const b64 = Buffer.from(key).toString('base64');
-
-	// Store in macOS Keychain (no user prompts with -U flag)
-	const add = Bun.spawn([
-		'security',
-		'add-generic-password',
-		'-s',
-		service,
-		'-a',
-		KEY_ACCOUNT,
-		'-w',
-		b64,
-		'-U', // Update without user confirmation
-	]);
-	await add.exited;
-
-	return key;
-}
-
-/**
- * Encrypt data using AES-256-GCM
- */
-async function encrypt(data: string, keyBytes: Uint8Array): Promise<Uint8Array> {
-	const key = await crypto.subtle.importKey('raw', keyBytes, 'AES-GCM', false, ['encrypt']);
-
-	const iv = crypto.getRandomValues(new Uint8Array(12));
-	const plaintext = new TextEncoder().encode(data);
-
-	const ciphertext = new Uint8Array(
-		await crypto.subtle.encrypt({ name: 'AES-GCM', iv }, key, plaintext)
-	);
-
-	// Combine IV + ciphertext
-	const combined = new Uint8Array(iv.length + ciphertext.length);
-	combined.set(iv, 0);
-	combined.set(ciphertext, iv.length);
-
-	return combined;
-}
-
-/**
- * Decrypt data using AES-256-GCM
- */
-async function decrypt(combined: Uint8Array, keyBytes: Uint8Array): Promise<string> {
-	const key = await crypto.subtle.importKey('raw', keyBytes, 'AES-GCM', false, ['decrypt']);
-
-	const iv = combined.slice(0, 12);
-	const ciphertext = combined.slice(12);
-
-	const plaintext = await crypto.subtle.decrypt({ name: 'AES-GCM', iv }, key, ciphertext);
-
-	return new TextDecoder().decode(plaintext);
-}
-
-async function saveEncryptedValueToKeychain(
-	service: string,
-	account: string,
-	value: string
-): Promise<void> {
-	const key = await ensureEncryptionKey(service);
-	const encrypted = await encrypt(value, key);
-	const b64 = Buffer.from(encrypted).toString('base64');
-
-	const del = Bun.spawn(['security', 'delete-generic-password', '-s', service, '-a', account], {
-		stderr: 'ignore',
-	});
-	await del.exited;
-
-	const add = Bun.spawn([
-		'security',
-		'add-generic-password',
-		'-s',
-		service,
-		'-a',
-		account,
-		'-w',
-		b64,
-		'-U',
-	]);
-	await add.exited;
-}
-
-async function getEncryptedValueFromKeychain(
-	service: string,
-	account: string
-): Promise<string | null> {
-	const find = Bun.spawn(
-		['security', 'find-generic-password', '-s', service, '-a', account, '-w'],
-		{ stderr: 'ignore' }
-	);
-
-	const stdout = await new Response(find.stdout).text();
-	if (stdout.length === 0) {
-		return null;
-	}
-
-	const encrypted = Uint8Array.from(Buffer.from(stdout.trim(), 'base64'));
-	const key = await ensureEncryptionKey(service);
-	return decrypt(encrypted, key);
-}
-
-async function deleteValueFromKeychain(service: string, account: string): Promise<void> {
-	const del = Bun.spawn(['security', 'delete-generic-password', '-s', service, '-a', account], {
-		stderr: 'ignore',
-	});
-	await del.exited;
-}
-
-/**
- * Store auth data in macOS Keychain
- */
-export async function saveAuthToKeychain(
-	profileName: string,
-	authData: { api_key: string; user_id: string; expires: number }
-): Promise<void> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-	await saveEncryptedValueToKeychain(service, AUTH_ACCOUNT, JSON.stringify(authData));
-}
-
-/**
- * Retrieve auth data from macOS Keychain
- */
-export async function getAuthFromKeychain(
-	profileName: string
-): Promise<{ api_key: string; user_id: string; expires: number } | null> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-
-	try {
-		const json = await getEncryptedValueFromKeychain(service, AUTH_ACCOUNT);
-		if (!json) {
-			return null;
-		}
-		return JSON.parse(json);
-	} catch {
-		return null;
-	}
-}
-
-/**
- * Delete auth data from macOS Keychain
- */
-export async function deleteAuthFromKeychain(profileName: string): Promise<void> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-	await deleteValueFromKeychain(service, AUTH_ACCOUNT);
-}
-
-export async function saveCoderApiKeyToKeychain(
-	profileName: string,
-	apiKey: string
-): Promise<void> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-	await saveEncryptedValueToKeychain(service, CODER_API_KEY_ACCOUNT, apiKey);
-}
-
-export async function getCoderApiKeyFromKeychain(profileName: string): Promise<string | null> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-	try {
-		return await getEncryptedValueFromKeychain(service, CODER_API_KEY_ACCOUNT);
-	} catch {
-		return null;
-	}
-}
-
-export async function deleteCoderApiKeyFromKeychain(profileName: string): Promise<void> {
-	const service = `${SERVICE_PREFIX}.${profileName}`;
-	await deleteValueFromKeychain(service, CODER_API_KEY_ACCOUNT);
-}

package/src/log-collector.ts

@@ -1,77 +0,0 @@
-/**
- * Log Collector for clean build logs
- *
- * Provides a mechanism to collect clean, non-animated log output
- * for streaming to external services (like Pulse) while keeping
- * animated TUI output for the user's terminal.
- *
- * Usage:
- * - Set AGENTUITY_CLEAN_LOGS_FILE env var to a file path
- * - TUI components call appendLog() for final state messages
- * - Logs are written to the file for the parent process to read
- */
-
-import { appendFileSync, writeFileSync } from 'node:fs';
-
-/**
- * Get the clean logs file path from environment
- */
-function getCleanLogsFile(): string | undefined {
-	return process.env.AGENTUITY_CLEAN_LOGS_FILE;
-}
-
-/**
- * Disable log collection (called on write errors to prevent repeated failures)
- */
-function disableLogCollection(error: unknown): void {
-	console.debug('Log collection disabled due to write error: %s', error);
-	delete process.env.AGENTUITY_CLEAN_LOGS_FILE;
-}
-
-/**
- * Check if log collection is enabled (via environment variable)
- */
-export function isLogCollectionEnabled(): boolean {
-	return !!getCleanLogsFile();
-}
-
-/**
- * Initialize the clean logs file (clears any existing content)
- */
-export function initCleanLogsFile(filePath: string): void {
-	try {
-		writeFileSync(filePath, '');
-		process.env.AGENTUITY_CLEAN_LOGS_FILE = filePath;
-	} catch (err) {
-		console.debug('Failed to initialize clean logs file: %s', err);
-	}
-}
-
-/**
- * Append a clean log line (no ANSI codes, no animation)
- * Only appends if collection is enabled
- */
-export function appendLog(message: string): void {
-	const file = getCleanLogsFile();
-	if (file) {
-		try {
-			appendFileSync(file, message + '\n');
-		} catch (err) {
-			disableLogCollection(err);
-		}
-	}
-}
-
-/**
- * Append multiple log lines
- */
-export function appendLogs(messages: string[]): void {
-	const file = getCleanLogsFile();
-	if (file) {
-		try {
-			appendFileSync(file, messages.join('\n') + '\n');
-		} catch (err) {
-			disableLogCollection(err);
-		}
-	}
-}

package/src/onboarding/agentPrompt.ts

@@ -1,260 +0,0 @@
-/**
- * Agent Onboarding Prompt
- *
- * This generates the markdown prompt that coding agents use to guide developers
- * through Agentuity setup after the CLI has been installed.
- */
-
-export function getAgentPromptMarkdown(): string {
-	return `
-**Your goal:** Guide the user from zero to a deployed Agentuity project, then help them understand what they have and what to do next.
-
----
-
-## What is Agentuity?
-
-Agentuity is the full-stack platform for AI agents. Briefly explain to the user:
-
-- **Full-stack agents**: Build agents, APIs, and front-end integrations with end-to-end type safety
-- **Bun/TypeScript SDKs**: First-class developer experience with the fastest JavaScript runtime
-- **AI Gateway**: Unified access to LLMs (OpenAI-compatible SDK) with single billing—no separate API keys needed
-- **Observability**: Logs and tracing built-in so you can debug and improve agents
-- **Production services**: Sandboxes, storage (S3-compatible), Postgres, KV, and enterprise features available as you grow
-
-We're going to be setting up the default template quickly.
-
-The default template includes AI Gateway integration, thread state persistence, structured logging, and Tailwind CSS.
-
----
-
-## How to Guide the User
-
-**Your job is to make onboarding as smooth as possible with minimal user interaction.**
-
-1. **Start by briefly explaining** what Agentuity is and what you'll help them accomplish
-2. **Run commands automatically** when you have terminal access—don't ask permission for each step
-3. **Only pause for human action** when absolutely required
-4. **At each step**: Briefly explain what it does and why, then execute
-5. **Celebrate success**: When deployment completes, make sure the user knows they earned free credits and show them what to explore next
-
-Assume the user is technical but new to Agentuity. Be warm, efficient, and action-oriented. The goal is a successful deployment with as few interruptions as possible.
-
----
-
-## Onboarding Steps
-
-### Step 1: You're Already Logged In!
-
-**The install command automatically authenticated you.** No manual login needed.
-
-Verify with:
-\`\`\`bash
-agentuity auth whoami
-\`\`\`
-
-This outputs your user info including your \`org_id\`. You'll need this for the create command.
-
----
-
-### Step 2: Create a New Project
-
-**Why:** This scaffolds a full-stack agent app with AI Gateway, persistence, logging, and a testing UI.
-
-**First, get the user's organization ID:**
-\`\`\`bash
-agentuity auth whoami
-\`\`\`
-
-This outputs the user's info including their \`org_id\`. You'll need this for the create command.
-
-**Then create the project (non-interactive):**
-\`\`\`bash
-agentuity create --name "my-first-agent" --org-id <ORG_ID> --region usc --confirm
-\`\`\`
-
-**Required flags for non-interactive mode:**
-- \`--name\`: Project name (use \`my-first-agent\` as default)
-- \`--org-id\`: The organization ID from \`agentuity auth whoami\` (required in non-interactive mode)
-- \`--region\`: Cloud region (use \`usc\` for US Central, or \`use\` for US East)
-- \`--confirm\`: Skips interactive prompts
-
-**Example with real org ID:**
-\`\`\`bash
-agentuity create --name "my-first-agent" --org-id ..... --region usc --confirm
-\`\`\`
-
-If the user wants a different location, add \`--dir\`:
-\`\`\`bash
-agentuity create --name "my-first-agent" --org-id <ORG_ID> --region usc --dir ~/projects/my-first-agent --confirm
-\`\`\`
-
-After creation, \`cd\` into the project directory.
-
-**Docs:** https://agentuity.dev/Get-Started/quickstart
-
----
-
-### Step 3: Run Locally
-
-**Why:** Start the development server to test the project locally before deploying.
-
-**Command:**
-\`\`\`bash
-agentuity dev
-\`\`\`
-
-**⚠️ This is a long-running process!** The dev server runs continuously until killed. Before or after starting it, **immediately tell the user what to do** — don't just wait silently.
-
-**What happens:**
-- **Local server** starts at \`http://localhost:3500\` — this is where your APIs and agents are served
-- **Public tunnel URL** is automatically created — the CLI will display a public URL that tunnels to your local server, useful for testing webhooks or sharing with others
-- **Hot reload** enabled — code changes reflect immediately
-
-**Immediately after starting the dev server, tell the user:**
-
-> "The dev server is now running! Here's what to try:
->
-> 1. Open **http://localhost:3500** to see the app
-> 2. The public tunnel URL shown in the terminal can be shared with others
->
-> Take a few minutes to explore. When you're ready to deploy to production, let me know!"
-
-**Wait for the user to confirm** they're ready to deploy before proceeding to Step 4.
-
-**Keyboard shortcuts** while dev server runs: \`h\` for help, \`r\` to restart, \`q\` to quit.
-
-**Docs:** https://agentuity.dev/Reference/CLI/development
-
----
-
-### Step 4: Deploy to Production
-
-**Why:** Make the project live with observability, logs, and analytics.
-
-**Important:** Deploying their first project earns the user **free credits** to continue experimenting with the platform. Make sure they know this is a reward for completing onboarding!
-
-**First, stop the dev server:**
-Before deploying, you must stop the dev server that's still running from Step 3. Send \`Ctrl+C\` or \`q\` to the terminal running \`agentuity dev\` to gracefully shut it down. Don't leave it running or you'll have an orphaned process on port 3500.
-
-**Then deploy:**
-\`\`\`bash
-agentuity deploy
-\`\`\`
-
-**What happens:**
-- Project builds and deploys to Agentuity's global edge network
-- The CLI will output the production URL when deployment succeeds
-- Observability (logs, tracing, analytics) is now live
-- **Free credits are awarded** for their first deployment
-
-**⚠️ CRITICAL: Do not hallucinate URLs!**
-- **Only show the user URLs that the CLI actually outputs**
-- Do not make up or guess production URLs (e.g., don't invent \`https://project-name.agentuity.app\`)
-- If the deploy command fails or doesn't show a URL, tell the user to check the dashboard instead
-- If there's an error, report the actual error message from the CLI
-
-**After successful deployment:**
-1. Show the user the **exact production URL from the CLI output**
-2. Direct them to the Agentuity dashboard at https://app.agentuity.com to:
-   - View their deployed project
-   - Explore logs and traces
-   - See their credit balance
-   - Monitor analytics
-
-**If deployment fails or shows errors:**
-- Report the actual error message
-- Don't claim success if the CLI didn't confirm it
-- Suggest checking the dashboard for deployment status
-
----
-
-## What the Template Provides
-
-After deployment, explain what they now have:
-
-| Feature | What It Does |
-|---------|--------------|
-| **AI Gateway** | OpenAI-compatible SDK routed through Agentuity—unified billing, no separate API keys |
-| **Thread State** | Conversations and agent state persist automatically for multi-turn flows |
-| **Structured Logging** | Logs and traces captured for debugging and analytics |
-| **Tailwind CSS** | Pre-configured styling for the frontend |
-| **Type Safety** | End-to-end types from agent schemas to API routes to frontend |
-
----
-
-## Next Steps
-
-After successful deployment, guide the user through these recommendations:
-
-### 1. Celebrate the Free Credits!
-- Remind the user: their first deployment earned them **free credits** to keep experimenting
-- They can view their credit balance in the Agentuity dashboard
-- These credits let them continue building and testing without upfront cost
-
-### 2. Join the Community
-- **Discord**: https://discord.gg/agentuity
-- Get help, see examples, share what you're building
-
-### 3. Explore the Dashboard
-- **Observability**: View logs and traces for your deployment
-- **Deployments**: See version history with rollback options
-- **Analytics**: Monitor usage and performance
-
-### 4. Connect GitHub
-- Enable GitHub integration in project settings for:
-  - Automatic deployments from branches
-  - CI/CD workflows
-  - Team collaboration
-
-### 5. Explore Additional Services
-As your project grows, Agentuity provides:
-- **Postgres**: Relational database
-- **KV**: Fast key-value storage
-- **Storage**: S3-compatible file storage
-- **Sandboxes**: Isolated execution environments
-
-### 6. Customize Your Agents
-The SDK makes it easy to:
-- Add new agents in \`src/agent/\` (just add a folder with \`agent.ts\`)
-- Create API routes in \`src/api/\`
-- Build type-safe frontend integrations
-
-**Docs:** https://agentuity.dev/Get-Started/project-structure
-
----
-
-## Handling Common Issues
-
-- **Bun not installed**: The CLI will prompt to install it, or run: \`curl -fsSL https://bun.sh/install | bash\`
-- **Login fails/times out**: Rerun \`agentuity login\` and check browser
-- **Port 3500 in use**: Kill the existing process or use \`agentuity dev --port 8080\`
-- **Deploy fails**: Check build errors in CLI output, ensure you're in the project directory
-
----
-
-## Success Criteria
-
-**Onboarding is successful when the user has:**
-
-1. ✅ **CLI installed** and authenticated
-2. ✅ **Local dev environment running** — they've seen \`localhost:3500\`
-3. ✅ **First deployment complete** — they have a live production URL
-4. ✅ **Free credits earned** — they know they've been rewarded and can keep building
-5. ✅ **Clear next steps** — they know about Discord, the dashboard, and how to extend their project
-
-**Make them feel accomplished!** They just went from zero to a deployed AI agent platform in minutes.
-
----
-
-## Style Guidelines
-
-- Keep explanations short and practical
-- Explain what & why at each step
-- Give exact commands
-- Handle errors calmly with concrete suggestions
-- Be explicit when human action is required
-- Use non-interactive CLI flags (\`--confirm\`, \`--name\`, \`--org-id\`, \`--region usc\`, \`--dir\`) for smooth automation
-- **Never hallucinate or make up information** — only report URLs, IDs, and status from actual CLI output
-- If something fails or you're unsure, say so honestly and direct the user to the dashboard
-`;
-}