xenfra 0.4.3__py3-none-any.whl → 0.4.5__py3-none-any.whl

xenfra/commands/deployments.py

@@ -1,973 +1,1358 @@
-"""
-Deployment commands for Xenfra CLI.
-"""
-
-import os
-
-import click
-from rich.console import Console
-from rich.panel import Panel
-from rich.table import Table
-from xenfra_sdk import XenfraClient
-from xenfra_sdk.exceptions import XenfraAPIError, XenfraError
-from xenfra_sdk.privacy import scrub_logs
-
-from ..utils.auth import API_BASE_URL, get_auth_token
-from ..utils.codebase import has_xenfra_config
-from ..utils.config import apply_patch, read_xenfra_yaml
-from ..utils.validation import (
-    validate_branch_name,
-    validate_deployment_id,
-    validate_framework,
-    validate_git_repo_url,
-    validate_project_name,
-)
-
-import time
-from datetime import datetime
-
-from rich.live import Live
-from rich.progress import BarColumn, Progress, SpinnerColumn, TextColumn
-
-console = Console()
-
-# Maximum number of retry attempts for auto-healing
-MAX_RETRY_ATTEMPTS = 3
-
-
-def get_client() -> XenfraClient:
-    """Get authenticated SDK client."""
-    token = get_auth_token()
-    if not token:
-        console.print("[bold red]Not logged in. Run 'xenfra auth login' first.[/bold red]")
-        raise click.Abort()
-
-    return XenfraClient(token=token, api_url=API_BASE_URL)
-
-
-def show_diagnosis_panel(diagnosis: str, suggestion: str):
-    """Display diagnosis and suggestion in formatted panels."""
-    console.print()
-    console.print(Panel(diagnosis, title="[bold red]🔍 Diagnosis[/bold red]", border_style="red"))
-    console.print()
-    console.print(
-        Panel(suggestion, title="[bold yellow]💡 Suggestion[/bold yellow]", border_style="yellow")
-    )
-
-
-def show_patch_preview(patch_data: dict):
-    """Show a preview of the patch that will be applied."""
-    console.print()
-    console.print("[bold green]🔧 Automatic Fix Available[/bold green]")
-    console.print(f"  [cyan]File:[/cyan] {patch_data.get('file')}")
-    console.print(f"  [cyan]Operation:[/cyan] {patch_data.get('operation')}")
-    console.print(f"  [cyan]Value:[/cyan] {patch_data.get('value')}")
-    console.print()
-
-
-def _stream_deployment(client: XenfraClient, project_name: str, git_repo: str, branch: str, framework: str, region: str, size_slug: str, is_dockerized: bool = True, port: int = None, command: str = None, entrypoint: str = None, database: str = None, package_manager: str = None, dependency_file: str = None, file_manifest: list = None):
-    """
-    Creates deployment with real-time SSE streaming (no polling needed).
-
-    Returns tuple of (status, deployment_id, logs_buffer)
-    """
-    console.print(Panel(
-        f"[bold cyan]Project:[/bold cyan] {project_name}\n"
-        f"[bold cyan]Mode:[/bold cyan] Real-time Streaming Deployment",
-        title="[bold green]🚀 Deployment Starting[/bold green]",
-        border_style="green"
-    ))
-
-    deployment_id = None
-    logs_buffer = []
-    status_val = "PENDING"
-
-    try:
-        for event in client.deployments.create_stream(
-            project_name=project_name,
-            git_repo=git_repo,
-            branch=branch,
-            framework=framework,
-            region=region,
-            size_slug=size_slug,
-            is_dockerized=is_dockerized,
-            port=port,
-            command=command,
-            entrypoint=entrypoint,  # Pass entrypoint to deployment API
-            database=database,
-            package_manager=package_manager,
-            dependency_file=dependency_file,
-            file_manifest=file_manifest,
-        ):
-
-            event_type = event.get("event", "message")
-            data = event.get("data", "")
-
-            if event_type == "deployment_created":
-                # Extract deployment ID
-                if isinstance(data, dict):
-                    deployment_id = data.get("deployment_id")
-                    console.print(f"[bold green]✓[/bold green] Deployment created: [cyan]{deployment_id}[/cyan]\n")
-
-            elif event_type == "log":
-                # Real-time log output
-                log_line = str(data)
-                logs_buffer.append(log_line)
-
-                # Colorize output
-                if any(x in log_line for x in ["ERROR", "FAILED", "✗"]):
-                    console.print(f"[bold red]{log_line}[/bold red]")
-                elif any(x in log_line for x in ["WARN", "WARNING", "⚠"]):
-                    console.print(f"[yellow]{log_line}[/yellow]")
-                elif any(x in log_line for x in ["SUCCESS", "COMPLETED", "✓", "passed!"]):
-                    console.print(f"[bold green]{log_line}[/bold green]")
-                elif "PHASE" in log_line:
-                    console.print(f"\n[bold blue]{log_line}[/bold blue]")
-                elif "[InfraEngine]" in log_line or "[INFO]" in log_line:
-                    console.print(f"[cyan]›[/cyan] {log_line}")
-                else:
-                    console.print(f"[dim]{log_line}[/dim]")
-
-            elif event_type == "error":
-                error_msg = str(data)
-                logs_buffer.append(f"ERROR: {error_msg}")
-                console.print(f"\n[bold red]❌ Error: {error_msg}[/bold red]")
-                status_val = "FAILED"
-
-            elif event_type == "deployment_complete":
-                # Final status
-                if isinstance(data, dict):
-                    status_val = data.get("status", "UNKNOWN")
-                    ip_address = data.get("ip_address")
-
-                    console.print()
-                    if status_val == "SUCCESS":
-                        console.print("[bold green]✨ SUCCESS: Your application is live![/bold green]")
-                        if ip_address and ip_address != "unknown":
-                            console.print(f"[bold]Accessible at:[/bold] [link=http://{ip_address}]http://{ip_address}[/link]")
-                    elif status_val == "FAILED":
-                        console.print("[bold red]❌ DEPLOYMENT FAILED[/bold red]")
-                        error = data.get("error")
-                        if error:
-                            console.print(f"[red]Error: {error}[/red]")
-                break
-
-    except Exception as e:
-        console.print(f"\n[bold red]❌ Streaming error: {e}[/bold red]")
-        status_val = "FAILED"
-        logs_buffer.append(f"Streaming error: {e}")
-
-    return (status_val, deployment_id, "\n".join(logs_buffer))
-
-
-def _follow_deployment(client: XenfraClient, deployment_id: str):
-    """
-    Polls logs and status in real-time until completion with CI/CD style output.
-    (LEGACY - Used for backward compatibility)
-    """
-    console.print(Panel(
-        f"[bold cyan]Deployment ID:[/bold cyan] {deployment_id}\n"
-        f"[bold cyan]Mode:[/bold cyan] Streaming Real-time Infrastructure Logs",
-        title="[bold green]🚀 Deployment Monitor[/bold green]",
-        border_style="green"
-    ))
-
-    last_log_len = 0
-    status_val = "PENDING"
-
-    # Use a live display for the progress bar at the bottom
-    with Progress(
-        SpinnerColumn(),
-        TextColumn("[bold blue]{task.description}"),
-        BarColumn(bar_width=40),
-        TextColumn("[progress.percentage]{task.percentage:>3.0f}%"),
-        console=console,
-        transient=False,
-    ) as progress:
-        task = progress.add_task("Waiting for server response...", total=100)
-
-        while status_val not in ["SUCCESS", "FAILED", "CANCELLED"]:
-            try:
-                # 1. Update Status
-                dep_status = client.deployments.get_status(deployment_id)
-                status_val = dep_status.get("status", "PENDING")
-                progress_val = dep_status.get("progress", 0)
-                state = dep_status.get("state", "preparing")
-
-                # Use a more descriptive description for the progress task
-                desc = f"Phase: {state}"
-                if status_val == "FAILED":
-                    desc = "[bold red]FAILED[/bold red]"
-                elif status_val == "SUCCESS":
-                    desc = "[bold green]SUCCESS[/bold green]"
-
-                progress.update(task, completed=progress_val, description=desc)
-
-                # 2. Update Logs
-                log_content = client.deployments.get_logs(deployment_id)
-                if log_content and len(log_content) > last_log_len:
-                    new_logs = log_content[last_log_len:].strip()
-                    for line in new_logs.split("\n"):
-                        # Process and colorize lines
-                        clean_line = line.strip()
-                        if not clean_line:
-                            continue
-
-                        if any(x in clean_line for x in ["ERROR", "FAILED", "✗"]):
-                            progress.console.print(f"[bold red]{clean_line}[/bold red]")
-                        elif any(x in clean_line for x in ["WARN", "WARNING", "⚠"]):
-                            progress.console.print(f"[yellow]{clean_line}[/yellow]")
-                        elif any(x in clean_line for x in ["SUCCESS", "COMPLETED", "✓", "passed!"]):
-                            progress.console.print(f"[bold green]{clean_line}[/bold green]")
-                        elif "PHASE" in clean_line:
-                            progress.console.print(f"\n[bold blue]{clean_line}[/bold blue]")
-                        elif "[InfraEngine]" in clean_line:
-                            progress.console.print(f"[dim]{clean_line}[/dim]")
-                        else:
-                            progress.console.print(f"[cyan]›[/cyan] {clean_line}")
-
-                    last_log_len = len(log_content)
-
-                if status_val in ["SUCCESS", "FAILED", "CANCELLED"]:
-                    break
-
-                time.sleep(1.5)  # Slightly faster polling for better feel
-            except Exception as e:
-                # progress.console.print(f"[dim]Transient connection issue: {e}[/dim]")
-                time.sleep(3)
-                continue
-
-    console.print()
-    if status_val == "SUCCESS":
-        console.print("[bold green]✨ SUCCESS: Your application is live![/bold green]")
-        # Try to get the IP address
-        try:
-            final_status = client.deployments.get_status(deployment_id)
-            ip = final_status.get("ip_address")
-            if ip:
-                console.print(f"[bold]Accessible at:[/bold] [link=http://{ip}]http://{ip}[/link]")
-        except:
-            pass
-    elif status_val == "FAILED":
-        console.print("\n[bold red]❌ FAILURE DETECTED: Entering AI Diagnosis Mode...[/bold red]")
-
-    return status_val
-
-
-def zen_nod_workflow(logs: str, client: XenfraClient, attempt: int) -> bool:
-    """
-    Execute the Zen Nod auto-healing workflow.
-
-    Args:
-        logs: Deployment error logs
-        client: Authenticated SDK client
-        attempt: Current attempt number
-
-    Returns:
-        True if patch was applied and user wants to retry, False otherwise
-    """
-    console.print()
-    console.print(f"[cyan]🤖 Analyzing failure (attempt {attempt}/{MAX_RETRY_ATTEMPTS})...[/cyan]")
-
-    # Slice logs to last 300 lines for focused diagnosis (Fix #26)
-    log_lines = logs.split("\n")
-    if len(log_lines) > 300:
-        logs = "\n".join(log_lines[-300:])
-        console.print("[dim]Note: Analyzing only the last 300 lines of logs for efficiency.[/dim]")
-
-    # Scrub sensitive data from logs
-    scrubbed_logs = scrub_logs(logs)
-
-    # Diagnose with AI
-    try:
-        diagnosis_result = client.intelligence.diagnose(scrubbed_logs)
-    except Exception as e:
-        console.print(f"[yellow]Could not diagnose failure: {e}[/yellow]")
-        return False
-
-    # Show diagnosis
-    show_diagnosis_panel(diagnosis_result.diagnosis, diagnosis_result.suggestion)
-
-    # Check if there's an automatic patch
-    if diagnosis_result.patch and diagnosis_result.patch.file:
-        show_patch_preview(diagnosis_result.patch.model_dump())
-
-        # Zen Nod confirmation
-        if click.confirm("Apply this fix and retry deployment?", default=True):
-            try:
-                # Apply patch (with automatic backup)
-                backup_path = apply_patch(diagnosis_result.patch.model_dump())
-                console.print("[bold green]✓ Patch applied[/bold green]")
-                if backup_path:
-                    console.print(f"[dim]Backup saved: {backup_path}[/dim]")
-                return True  # Signal to retry
-            except Exception as e:
-                console.print(f"[bold red]Failed to apply patch: {e}[/bold red]")
-                return False
-        else:
-            console.print()
-            console.print("[yellow]❌ Patch declined. Follow the manual steps above.[/yellow]")
-            return False
-    else:
-        console.print()
-        console.print(
-            "[yellow]No automatic fix available. Please follow the manual steps above.[/yellow]"
-        )
-        return False
-
-
-@click.command()
-@click.option("--project-name", help="Project name (defaults to current directory name)")
-@click.option("--git-repo", help="Git repository URL (if deploying from git)")
-@click.option("--branch", default="main", help="Git branch (default: main)")
-@click.option("--framework", help="Framework override (fastapi, flask, django)")
-@click.option("--region", help="DigitalOcean region override")
-@click.option("--size", help="DigitalOcean size slug override")
-@click.option("--no-heal", is_flag=True, help="Disable auto-healing on failure")
-def deploy(project_name, git_repo, branch, framework, region, size, no_heal):
-    """
-    Deploy current project to DigitalOcean with auto-healing.
-
-    Deploys your application with zero configuration. The CLI will:
-    1. Check for xenfra.yaml (or run init if missing)
-    2. Create a deployment
-    3. Auto-diagnose and fix failures (unless --no-heal is set or XENFRA_NO_AI=1)
-
-    Set XENFRA_NO_AI=1 environment variable to disable all AI features.
-    """
-    # Check XENFRA_NO_AI environment variable
-    no_ai = os.environ.get("XENFRA_NO_AI", "0") == "1"
-    if no_ai:
-        console.print("[yellow]XENFRA_NO_AI is set. Auto-healing disabled.[/yellow]")
-        no_heal = True
-
-    # Check for xenfra.yaml
-    if not has_xenfra_config():
-        console.print("[yellow]No xenfra.yaml found.[/yellow]")
-        if click.confirm("Run 'xenfra init' to create configuration?", default=True):
-            from .intelligence import init
-
-            ctx = click.get_current_context()
-            ctx.invoke(init, manual=no_ai, accept_all=False)
-        else:
-            console.print("[dim]Deployment cancelled.[/dim]")
-            return
-
-    # Load configuration from xenfra.yaml if it exists
-    config = {}
-    if has_xenfra_config():
-        try:
-            config = read_xenfra_yaml()
-        except Exception as e:
-            console.print(f"[yellow]Warning: Could not read xenfra.yaml: {e}[/dim]")
-
-    # Resolve values with precedence: 1. CLI Flag, 2. xenfra.yaml, 3. Default
-    project_name = project_name or config.get("name") or os.path.basename(os.getcwd())
-    framework = framework or config.get("framework")
-    # Track if is_dockerized was explicitly set in config (to avoid AI override)
-    is_dockerized_from_config = "is_dockerized" in config
-    is_dockerized = config.get("is_dockerized", True)
-    region = region or config.get("region") or "nyc3"
-    
-    # Resolve size slug (complex mapping)
-    if not size:
-        if config.get("size"):
-            size = config.get("size")
-        else:
-            instance_size = config.get("instance_size", "basic")
-            resources = config.get("resources", {})
-            cpu = resources.get("cpu", 1)
-            
-            if instance_size == "standard" or cpu >= 2:
-                size = "s-2vcpu-4gb"
-            elif instance_size == "premium" or cpu >= 4:
-                size = "s-4vcpu-8gb"
-            else:
-                size = "s-1vcpu-1gb"
-
-    # Extract port, command, database from config
-    # Track if port was explicitly set to avoid AI override
-    port_from_config = config.get("port")
-    port = port_from_config or 8000
-    command = config.get("command")  # Auto-detected if not provided
-    entrypoint = config.get("entrypoint")  # e.g., "todo.main:app"
-    database_config = config.get("database", {})
-    database = database_config.get("type") if isinstance(database_config, dict) else None
-    package_manager = config.get("package_manager", "pip")
-    dependency_file = config.get("dependency_file", "requirements.txt")
-
-    # Default project name to current directory
-    if not project_name:
-        project_name = os.path.basename(os.getcwd())
-
-    # Validate project name
-    is_valid, error_msg = validate_project_name(project_name)
-    if not is_valid:
-        console.print(f"[bold red]Invalid project name: {error_msg}[/bold red]")
-        raise click.Abort()
-
-    # Validate git repo if provided
-    if git_repo:
-        is_valid, error_msg = validate_git_repo_url(git_repo)
-        if not is_valid:
-            console.print(f"[bold red]Invalid git repository URL: {error_msg}[/bold red]")
-            raise click.Abort()
-        console.print(f"[cyan]Deploying {project_name} from git repository...[/cyan]")
-        console.print(f"[dim]Repository: {git_repo} (branch: {branch})[/dim]")
-    else:
-        # Note: Local folder deployment only works when engine runs locally
-        # In cloud API mode, this will fail with a clear error from the server
-        console.print(f"[cyan]Deploying {project_name}...[/cyan]")
-        console.print("[dim]Note: Git repository recommended for cloud deployments[/dim]")
-
-    # Validate branch name
-    is_valid, error_msg = validate_branch_name(branch)
-    if not is_valid:
-        console.print(f"[bold red]Invalid branch name: {error_msg}[/bold red]")
-        raise click.Abort()
-
-    # Validate framework if provided
-    if framework:
-        is_valid, error_msg = validate_framework(framework)
-        if not is_valid:
-            console.print(f"[bold red]Invalid framework: {error_msg}[/bold red]")
-            raise click.Abort()
-
-    # Retry loop for auto-healing
-    attempt = 0
-    deployment_id = None
-
-    try:
-        with get_client() as client:
-            while attempt < MAX_RETRY_ATTEMPTS:
-                attempt += 1
-
-                if attempt > 1:
-                    console.print(
-                        f"\n[cyan]🔄 Retrying deployment (attempt {attempt}/{MAX_RETRY_ATTEMPTS})...[/cyan]"
-                    )
-                else:
-                    console.print("[cyan]Creating deployment...[/cyan]")
-
-                # Detect framework if not provided (AI-powered Zen Mode)
-                if not framework:
-                    console.print("[cyan]🔍 AI Auto-detecting project type...[/cyan]")
-                    try:
-                        from ..utils.codebase import scan_codebase
-                        code_snippets = scan_codebase()
-                        if code_snippets:
-                            analysis = client.intelligence.analyze_codebase(code_snippets)
-                            framework = analysis.framework
-                            # Only use AI's is_dockerized if config didn't explicitly set it
-                            if not is_dockerized_from_config:
-                                is_dockerized = analysis.is_dockerized
-                            # Override port if AI detected it and config didn't set one
-                            if not port_from_config and analysis.port:
-                                port = analysis.port
-                            # Override port and size if AI has strong recommendations
-                            if not size and analysis.instance_size:
-                                size = "s-1vcpu-1gb" if analysis.instance_size == "basic" else "s-2vcpu-4gb"
-                            
-                            mode_str = "Docker" if is_dockerized else "Bare Metal"
-                            console.print(f"[green]✓ Detected {framework.upper()} project ({mode_str} Mode)[/green] (Port: {port})")
-                        else:
-                            console.print("[yellow]⚠ No code files found for AI analysis. Defaulting to 'fastapi'[/yellow]")
-                            framework = "fastapi"
-                            is_dockerized = True
-                    except Exception as e:
-                        console.print(f"[yellow]⚠ AI detection failed: {e}. Defaulting to 'fastapi'[/yellow]")
-                        framework = "fastapi"
-                        is_dockerized = True
-
-                # Delta upload: if no git_repo, scan and upload local files
-                file_manifest = None
-                if not git_repo:
-                    from ..utils.file_sync import scan_project_files_cached, ensure_gitignore_ignored
-                    
-                    # Protect privacy: ensure .xenfra is in .gitignore
-                    if ensure_gitignore_ignored():
-                        console.print("[dim]   - Added .xenfra to .gitignore for privacy[/dim]")
-                    
-                    console.print("[cyan]📁 Scanning project files...[/cyan]")
-
-                    file_manifest = scan_project_files_cached()
-                    console.print(f"[dim]Found {len(file_manifest)} files[/dim]")
-                    
-                    if not file_manifest:
-                        console.print("[bold red]Error: No files found to deploy.[/bold red]")
-                        raise click.Abort()
-                    
-                    # Check which files need uploading
-                    console.print("[cyan]🔍 Checking file cache...[/cyan]")
-                    check_result = client.files.check(file_manifest)
-                    missing = check_result.get('missing', [])
-                    cached = check_result.get('cached', 0)
-                    
-                    if cached > 0:
-                        console.print(f"[green]✓ {cached} files already cached[/green]")
-                    
-                    # Upload missing files
-                    if missing:
-                        console.print(f"[cyan]☁️ Uploading {len(missing)} files...[/cyan]")
-                        uploaded = client.files.upload_files(
-                            file_manifest,
-                            missing,
-                            progress_callback=lambda done, total: console.print(f"[dim]   Progress: {done}/{total}[/dim]") if done % 10 == 0 or done == total else None
-                        )
-                        console.print(f"[green]✓ Uploaded {uploaded} files[/green]")
-                    else:
-                        console.print("[green]✓ All files already cached[/green]")
-                    
-                    # Remove abs_path from manifest before sending to API
-                    file_manifest = [{"path": f["path"], "sha": f["sha"], "size": f["size"]} for f in file_manifest]
-
-                # Create deployment with real-time streaming
-                try:
-                    status_result, deployment_id, logs_data = _stream_deployment(
-                        client=client,
-                        project_name=project_name,
-                        git_repo=git_repo,
-                        branch=branch,
-                        framework=framework,
-                        region=region,
-                        size_slug=size,
-                        is_dockerized=is_dockerized,
-                        port=port,
-                        command=command,
-                        entrypoint=entrypoint,  # Pass entrypoint to deployment API
-                        database=database,
-                        package_manager=package_manager,
-                        dependency_file=dependency_file,
-                        file_manifest=file_manifest,
-                    )
-
-
-                    if status_result == "FAILED" and not no_heal:
-                        # Hand off to the Zen Nod AI Agent
-                        should_retry = zen_nod_workflow(logs_data, client, attempt)
-
-                        if should_retry:
-                            # The agent applied a fix, loop back for attempt + 1
-                            continue
-                        else:
-                            # Agent couldn't fix it or user declined
-                            raise click.Abort()
-
-                    # If we got here with success, break the retry loop
-                    if status_result == "SUCCESS":
-                        break
-                    else:
-                        raise click.Abort()
-
-                except XenfraAPIError as e:
-                    # Deployment failed - try to provide helpful error
-                    from ..utils.errors import detect_error_type, show_error_with_solution
-                    
-                    console.print(f"[bold red]✗ Deployment failed[/bold red]")
-                    
-                    # Try to detect error type and show solution
-                    error_type, error_kwargs = detect_error_type(str(e.detail))
-                    if error_type:
-                        show_error_with_solution(error_type, **error_kwargs)
-                    else:
-                        console.print(f"[red]{e.detail}[/red]")
-
-                    # Check if we should auto-heal
-                    if no_heal or attempt >= MAX_RETRY_ATTEMPTS:
-                        # No auto-healing or max retries reached
-                        if attempt >= MAX_RETRY_ATTEMPTS:
-                            console.print(
-                                f"\n[bold red]❌ Maximum retry attempts ({MAX_RETRY_ATTEMPTS}) reached.[/bold red]"
-                            )
-                            console.print(
-                                "[yellow]Unable to auto-fix the issue. Please review the errors above.[/yellow]"
-                            )
-                        raise
-                    else:
-                        # Try to get logs for diagnosis
-                        error_logs = str(e.detail)
-                        try:
-                            if deployment_id:
-                                # This should be a method in the SDK that returns a string
-                                logs_response = client.deployments.get_logs(deployment_id)
-                                if isinstance(logs_response, dict):
-                                    error_logs = logs_response.get("logs", str(e.detail))
-                                else:
-                                    error_logs = str(logs_response)  # Assuming it can be a string
-                        except Exception as log_err:
-                            console.print(
-                                f"[yellow]Warning: Could not fetch detailed logs for diagnosis: {log_err}[/yellow]"
-                            )
-                            # Fallback to the initial error detail
-                            pass
-
-                        # Run Zen Nod workflow
-                        should_retry = zen_nod_workflow(error_logs, client, attempt)
-
-                        if not should_retry:
-                            # User declined patch or no patch available
-                            console.print("\n[dim]Deployment cancelled.[/dim]")
-                            raise click.Abort()
-
-                        # Continue to next iteration (retry)
-                        continue
-
-    except XenfraAPIError as e:
-        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
-    except XenfraError as e:
-        console.print(f"[bold red]Error: {e}[/bold red]")
-    except click.Abort:
-        pass
-    except Exception as e:
-        console.print(f"[bold red]Unexpected error: {e}[/bold red]")
-
-
-@click.command()
-@click.argument("deployment-id")
-@click.option("--follow", "-f", is_flag=True, help="Follow log output (stream)")
-@click.option("--tail", type=int, help="Show last N lines")
-def logs(deployment_id, follow, tail):
-    # Validate deployment ID
-    is_valid, error_msg = validate_deployment_id(deployment_id)
-    if not is_valid:
-        console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
-        raise click.Abort()
-    """
-    Stream deployment logs.
-
-    Shows logs for a specific deployment. Use --follow to stream logs in real-time.
-    """
-    try:
-        with get_client() as client:
-            console.print(f"[cyan]Fetching logs for deployment {deployment_id}...[/cyan]")
-
-            log_content = client.deployments.get_logs(deployment_id)
-
-            if not log_content:
-                console.print("[yellow]No logs available yet.[/yellow]")
-                console.print("[dim]The deployment may still be starting up.[/dim]")
-                return
-
-            # Process logs
-            log_lines = log_content.strip().split("\n")
-
-            # Apply tail if specified
-            if tail:
-                log_lines = log_lines[-tail:]
-
-            # Display logs with syntax highlighting
-            console.print(f"\n[bold]Logs for deployment {deployment_id}:[/bold]\n")
-
-            if follow:
-                _follow_deployment(client, deployment_id)
-                return
-
-            # Display logs
-            for line in log_lines:
-                # Color-code based on log level
-                if "ERROR" in line or "FAILED" in line:
-                    console.print(f"[red]{line}[/red]")
-                elif "WARN" in line or "WARNING" in line:
-                    console.print(f"[yellow]{line}[/yellow]")
-                elif "SUCCESS" in line or "COMPLETED" in line:
-                    console.print(f"[green]{line}[/green]")
-                elif "INFO" in line:
-                    console.print(f"[cyan]{line}[/cyan]")
-                else:
-                    console.print(line)
-
-    except XenfraAPIError as e:
-        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
-    except XenfraError as e:
-        console.print(f"[bold red]Error: {e}[/bold red]")
-    except click.Abort:
-        pass
-
-
-@click.command()
-@click.argument("deployment-id", required=False)
-@click.option("--watch", "-w", is_flag=True, help="Watch status updates")
-def status(deployment_id, watch):
-    """
-    Show deployment status.
-
-    Displays current status, progress, and details for a deployment.
-    Use --watch to monitor status in real-time.
-    """
-    try:
-        if not deployment_id:
-            console.print("[yellow]No deployment ID provided.[/yellow]")
-            console.print("[dim]Usage: xenfra status <deployment-id>[/dim]")
-            return
-
-        # Validate deployment ID
-        is_valid, error_msg = validate_deployment_id(deployment_id)
-        if not is_valid:
-            console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
-            raise click.Abort()
-
-        with get_client() as client:
-            console.print(f"[cyan]Fetching status for deployment {deployment_id}...[/cyan]")
-
-            deployment_status = client.deployments.get_status(deployment_id)
-
-            if watch:
-                _follow_deployment(client, deployment_id)
-                return
-
-            # Display status
-            status_value = deployment_status.get("status", "UNKNOWN")
-            state = deployment_status.get("state", "unknown")
-            progress = deployment_status.get("progress", 0)
-
-            # Status panel
-            status_color = {
-                "PENDING": "yellow",
-                "IN_PROGRESS": "cyan",
-                "SUCCESS": "green",
-                "FAILED": "red",
-                "CANCELLED": "dim",
-            }.get(status_value, "white")
-
-            # Create status table
-            table = Table(show_header=False, box=None)
-            table.add_column("Property", style="cyan")
-            table.add_column("Value")
-
-            table.add_row("Deployment ID", str(deployment_id))
-            table.add_row("Status", f"[{status_color}]{status_value}[/{status_color}]")
-            table.add_row("State", state)
-
-            if progress > 0:
-                table.add_row("Progress", f"{progress}%")
-
-            if "project_name" in deployment_status:
-                table.add_row("Project", deployment_status["project_name"])
-
-            if "created_at" in deployment_status:
-                table.add_row("Created", deployment_status["created_at"])
-
-            if "finished_at" in deployment_status:
-                table.add_row("Finished", deployment_status["finished_at"])
-
-            if "url" in deployment_status:
-                table.add_row("URL", f"[link]{deployment_status['url']}[/link]")
-
-            if "ip_address" in deployment_status:
-                table.add_row("IP Address", deployment_status["ip_address"])
-
-            panel = Panel(table, title="[bold]Deployment Status[/bold]", border_style=status_color)
-            console.print(panel)
-
-            # Show error if failed
-            if status_value == "FAILED" and "error" in deployment_status:
-                error_panel = Panel(
-                    deployment_status["error"],
-                    title="[bold red]Error[/bold red]",
-                    border_style="red",
-                )
-                console.print("\n", error_panel)
-
-                console.print("\n[bold]Troubleshooting:[/bold]")
-                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
-                console.print(f"  • Diagnose: [cyan]xenfra diagnose {deployment_id}[/cyan]")
-
-            # Show next steps based on status
-            elif status_value == "SUCCESS":
-                console.print("\n[bold green]Deployment successful! 🎉[/bold green]")
-                if "url" in deployment_status:
-                    console.print(f"  • Visit: [link]{deployment_status['url']}[/link]")
-
-            elif status_value in ["PENDING", "IN_PROGRESS"]:
-                console.print("\n[bold]Deployment in progress...[/bold]")
-                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
-                console.print(f"  • Check again: [cyan]xenfra status {deployment_id}[/cyan]")
-
-    except XenfraAPIError as e:
-        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
-    except XenfraError as e:
-        console.print(f"[bold red]Error: {e}[/bold red]")
-    except click.Abort:
-        pass
-
-
-@click.command()
-@click.argument("deployment-id")
-@click.option("--format", "output_format", type=click.Choice(["detailed", "summary"], case_sensitive=False), default="detailed", help="Report format (detailed or summary)")
-def report(deployment_id, output_format):
-    """
-    Generate deployment report with self-healing events.
-    
-    Shows comprehensive deployment information including:
-    - Deployment status and timeline
-    - Self-healing attempts and outcomes
-    - Patches applied during healing
-    - Statistics and metrics
-    """
-    try:
-        # Validate deployment ID
-        is_valid, error_msg = validate_deployment_id(deployment_id)
-        if not is_valid:
-            console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
-            raise click.Abort()
-
-        with get_client() as client:
-            console.print(f"[cyan]Generating report for deployment {deployment_id}...[/cyan]\n")
-
-            # Get deployment status
-            try:
-                deployment_status = client.deployments.get_status(deployment_id)
-            except XenfraAPIError as e:
-                console.print(f"[bold red]Error fetching deployment status: {e.detail}[/bold red]")
-                raise click.Abort()
-
-            # Get deployment logs
-            try:
-                logs = client.deployments.get_logs(deployment_id)
-            except XenfraAPIError:
-                logs = None
-
-            # Parse status
-            status_value = deployment_status.get("status", "UNKNOWN")
-            state = deployment_status.get("state", "unknown")
-            progress = deployment_status.get("progress", 0)
-
-            # Status color mapping
-            status_color = {
-                "PENDING": "yellow",
-                "IN_PROGRESS": "cyan",
-                "SUCCESS": "green",
-                "FAILED": "red",
-                "CANCELLED": "dim",
-            }.get(status_value, "white")
-
-            # Calculate statistics from logs
-            heal_attempts = logs.count("🤖 Analyzing failure") if logs else 0
-            patches_applied = logs.count("✓ Patch applied") if logs else 0
-            diagnoses = logs.count("🔍 Diagnosis") if logs else 0
-
-            # Create main report table
-            report_table = Table(show_header=True, box=None)
-            report_table.add_column("Property", style="cyan", width=25)
-            report_table.add_column("Value", style="white")
-
-            report_table.add_row("Deployment ID", str(deployment_id))
-            report_table.add_row("Status", f"[{status_color}]{status_value}[/{status_color}]")
-            report_table.add_row("State", state)
-            
-            if progress > 0:
-                report_table.add_row("Progress", f"{progress}%")
-
-            if "project_name" in deployment_status:
-                report_table.add_row("Project", deployment_status["project_name"])
-
-            if "created_at" in deployment_status:
-                report_table.add_row("Created", deployment_status["created_at"])
-
-            if "finished_at" in deployment_status:
-                report_table.add_row("Finished", deployment_status["finished_at"])
-
-            if "url" in deployment_status:
-                report_table.add_row("URL", f"[link]{deployment_status['url']}[/link]")
-
-            if "ip_address" in deployment_status:
-                report_table.add_row("IP Address", deployment_status["ip_address"])
-
-            # Self-healing statistics
-            report_table.add_row("", "")  # Separator
-            report_table.add_row("[bold]Self-Healing Stats[/bold]", "")
-            report_table.add_row("Healing Attempts", str(heal_attempts))
-            report_table.add_row("Patches Applied", str(patches_applied))
-            report_table.add_row("Diagnoses Performed", str(diagnoses))
-            
-            if heal_attempts > 0:
-                success_rate = (patches_applied / heal_attempts * 100) if heal_attempts > 0 else 0
-                report_table.add_row("Healing Success Rate", f"{success_rate:.1f}%")
-
-            # Display main report
-            console.print(Panel(report_table, title="[bold]Deployment Report[/bold]", border_style=status_color))
-
-            # Detailed format includes timeline and healing events
-            if output_format == "detailed" and logs:
-                console.print("\n[bold]Self-Healing Timeline[/bold]\n")
-                
-                # Extract healing events from logs
-                log_lines = logs.split("\n")
-                timeline_entries = []
-                
-                for i, line in enumerate(log_lines):
-                    if "🤖 Analyzing failure" in line:
-                        attempt_match = None
-                        # Try to find attempt number in surrounding lines
-                        for j in range(max(0, i-5), min(len(log_lines), i+10)):
-                            if "attempt" in log_lines[j].lower():
-                                timeline_entries.append(("Healing Attempt", log_lines[j].strip()))
-                                break
-                    elif "🔍 Diagnosis" in line or "Diagnosis" in line:
-                        # Extract diagnosis text from next few lines
-                        diagnosis_text = line.strip()
-                        if i+1 < len(log_lines) and log_lines[i+1].strip():
-                            diagnosis_text += "\n  " + log_lines[i+1].strip()[:100]
-                        timeline_entries.append(("Diagnosis", diagnosis_text))
-                    elif "✓ Patch applied" in line or "Patch applied" in line:
-                        timeline_entries.append(("Patch Applied", line.strip()))
-                    elif "🔄 Retrying deployment" in line:
-                        timeline_entries.append(("Retry", line.strip()))
-
-                if timeline_entries:
-                    timeline_table = Table(show_header=True, box=None)
-                    timeline_table.add_column("Event", style="cyan", width=20)
-                    timeline_table.add_column("Details", style="white")
-
-                    for event_type, details in timeline_entries[:20]:  # Limit to 20 entries
-                        timeline_table.add_row(event_type, details)
-
-                    console.print(timeline_table)
-                else:
-                    console.print("[dim]No self-healing events detected in logs.[/dim]")
-
-            # Show error if failed
-            if status_value == "FAILED":
-                console.print("\n[bold red]⚠ Deployment Failed[/bold red]")
-                if "error" in deployment_status:
-                    error_panel = Panel(
-                        deployment_status["error"],
-                        title="[bold red]Error Details[/bold red]",
-                        border_style="red",
-                    )
-                    console.print("\n", error_panel)
-
-                console.print("\n[bold]Troubleshooting:[/bold]")
-                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
-                console.print(f"  • Diagnose: [cyan]xenfra diagnose {deployment_id}[/cyan]")
-                console.print(f"  • View status: [cyan]xenfra status {deployment_id}[/cyan]")
-
-            # Success summary
-            elif status_value == "SUCCESS":
-                console.print("\n[bold green]✓ Deployment Successful[/bold green]")
-                if heal_attempts > 0:
-                    console.print(f"[dim]Deployment succeeded after {heal_attempts} self-healing attempt(s).[/dim]")
-                if "url" in deployment_status:
-                    console.print(f"  • Visit: [link]{deployment_status['url']}[/link]")
-
-            # Export summary format (JSON-like structure for programmatic use)
-            if output_format == "summary":
-                console.print("\n[bold]Summary Format:[/bold]")
-                import json
-                summary = {
-                    "deployment_id": deployment_id,
-                    "status": status_value,
-                    "healing_attempts": heal_attempts,
-                    "patches_applied": patches_applied,
-                    "success": status_value == "SUCCESS",
-                }
-                console.print(f"[dim]{json.dumps(summary, indent=2)}[/dim]")
-
-    except XenfraAPIError as e:
-        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
-    except XenfraError as e:
-        console.print(f"[bold red]Error: {e}[/bold red]")
-    except click.Abort:
-        pass
-    except Exception as e:
-        console.print(f"[bold red]Unexpected error: {e}[/bold red]")
+"""
+Deployment commands for Xenfra CLI.
+"""
+
+import os
+from pathlib import Path
+
+import click
+from rich.console import Console
+from rich.panel import Panel
+from rich.table import Table
+from xenfra_sdk import XenfraClient
+from xenfra_sdk.exceptions import XenfraAPIError, XenfraError
+from xenfra_sdk.privacy import scrub_logs
+
+from ..utils.auth import API_BASE_URL, get_auth_token
+from ..utils.codebase import has_xenfra_config
+from ..utils.config import apply_patch, read_xenfra_yaml
+from ..blueprints import run_dry_run
+from ..utils.validation import (
+    validate_branch_name,
+    validate_deployment_id,
+    validate_framework,
+    validate_git_repo_url,
+    validate_project_id,
+    validate_project_name,
+)
+
+import time
+from datetime import datetime
+
+from rich.live import Live
+from rich.progress import BarColumn, Progress, SpinnerColumn, TextColumn
+
+console = Console()
+
+# Maximum number of retry attempts for auto-healing
+MAX_RETRY_ATTEMPTS = 3
+
+
+def get_client() -> XenfraClient:
+    """Get authenticated SDK client."""
+    token = get_auth_token()
+    if not token:
+        console.print("[bold red]Not logged in. Run 'xenfra auth login' first.[/bold red]")
+        raise click.Abort()
+
+    return XenfraClient(token=token, api_url=API_BASE_URL)
+
+
+def show_diagnosis_panel(diagnosis: str, suggestion: str):
+    """Display diagnosis and suggestion in formatted panels."""
+    console.print()
+    console.print(Panel(diagnosis, title="[bold red]🔍 Diagnosis[/bold red]", border_style="red"))
+    console.print()
+    console.print(
+        Panel(suggestion, title="[bold yellow]💡 Suggestion[/bold yellow]", border_style="yellow")
+    )
+
+
+def show_patch_preview(patch_data: dict):
+    """Show a preview of the patch that will be applied."""
+    console.print()
+    console.print("[bold green]🔧 Automatic Fix Available[/bold green]")
+    console.print(f"  [cyan]File:[/cyan] {patch_data.get('file')}")
+    console.print(f"  [cyan]Operation:[/cyan] {patch_data.get('operation')}")
+    console.print(f"  [cyan]Value:[/cyan] {patch_data.get('value')}")
+    console.print()
+
+
+def _stream_deployment(client: XenfraClient, project_name: str, git_repo: str, branch: str, framework: str, region: str, size_slug: str, is_dockerized: bool = True, port: int = None, command: str = None, entrypoint: str = None, database: str = None, package_manager: str = None, dependency_file: str = None, file_manifest: list = None, cleanup_on_failure: bool = False, services: list = None, mode: str = None):
+    """
+    Creates deployment with real-time SSE streaming (no polling needed).
+
+    Returns tuple of (status, deployment_id, logs_buffer)
+    """
+    console.print(Panel(
+        f"[bold cyan]Project:[/bold cyan] {project_name}\n"
+        f"[bold cyan]Mode:[/bold cyan] Real-time Streaming Deployment",
+        title="[bold green]🚀 Deployment Starting[/bold green]",
+        border_style="green"
+    ))
+
+    deployment_id = None
+    logs_buffer = []
+    status_val = "PENDING"
+
+    try:
+        for event in client.deployments.create_stream(
+            project_name=project_name,
+            git_repo=git_repo,
+            branch=branch,
+            framework=framework,
+            region=region,
+            size_slug=size_slug,
+            is_dockerized=is_dockerized,
+            port=port,
+            command=command,
+            entrypoint=entrypoint,  # Pass entrypoint to deployment API
+            database=database,
+            package_manager=package_manager,
+            dependency_file=dependency_file,
+            file_manifest=file_manifest,
+            cleanup_on_failure=cleanup_on_failure,
+            services=services,  # Microservices support
+            mode=mode,  # Deployment mode
+        ):
+
+            event_type = event.get("event", "message")
+            data = event.get("data", "")
+
+            if event_type == "deployment_created":
+                # Extract deployment ID
+                if isinstance(data, dict):
+                    deployment_id = data.get("deployment_id")
+                    console.print(f"[bold green]✓[/bold green] Deployment created: [cyan]{deployment_id}[/cyan]\n")
+
+            elif event_type == "log":
+                # Real-time log output
+                log_line = str(data)
+                logs_buffer.append(log_line)
+
+                # Colorize output
+                if any(x in log_line for x in ["ERROR", "FAILED", "✗"]):
+                    console.print(f"[bold red]{log_line}[/bold red]")
+                elif any(x in log_line for x in ["WARN", "WARNING", "⚠"]):
+                    console.print(f"[yellow]{log_line}[/yellow]")
+                elif any(x in log_line for x in ["SUCCESS", "COMPLETED", "✓", "passed!"]):
+                    console.print(f"[bold green]{log_line}[/bold green]")
+                elif "PHASE" in log_line:
+                    console.print(f"\n[bold blue]{log_line}[/bold blue]")
+                elif "[InfraEngine]" in log_line or "[INFO]" in log_line:
+                    console.print(f"[cyan]›[/cyan] {log_line}")
+                else:
+                    console.print(f"[dim]{log_line}[/dim]")
+
+            elif event_type == "error":
+                error_msg = str(data)
+                logs_buffer.append(f"ERROR: {error_msg}")
+                console.print(f"\n[bold red]❌ Error: {error_msg}[/bold red]")
+                status_val = "FAILED"
+
+            elif event_type == "deployment_complete":
+                # Final status
+                if isinstance(data, dict):
+                    status_val = data.get("status", "UNKNOWN")
+                    ip_address = data.get("ip_address")
+
+                    console.print()
+                    if status_val == "SUCCESS":
+                        console.print("[bold green]✨ SUCCESS: Your application is live![/bold green]")
+                        if ip_address and ip_address != "unknown":
+                            console.print(f"[bold]Accessible at:[/bold] [link=http://{ip_address}]http://{ip_address}[/link]")
+                    elif status_val == "FAILED":
+                        console.print("[bold red]❌ DEPLOYMENT FAILED[/bold red]")
+                        error = data.get("error")
+                        if error:
+                            console.print(f"[red]Error: {error}[/red]")
+                break
+
+    except Exception as e:
+        console.print(f"\n[bold red]❌ Streaming error: {e}[/bold red]")
+        status_val = "FAILED"
+        logs_buffer.append(f"Streaming error: {e}")
+
+    return (status_val, deployment_id, "\n".join(logs_buffer))
+
+
+def _follow_deployment(client: XenfraClient, deployment_id: str):
+    """
+    Polls logs and status in real-time until completion with CI/CD style output.
+    (LEGACY - Used for backward compatibility)
+    """
+    console.print(Panel(
+        f"[bold cyan]Deployment ID:[/bold cyan] {deployment_id}\n"
+        f"[bold cyan]Mode:[/bold cyan] Streaming Real-time Infrastructure Logs",
+        title="[bold green]🚀 Deployment Monitor[/bold green]",
+        border_style="green"
+    ))
+
+    last_log_len = 0
+    status_val = "PENDING"
+
+    # Use a live display for the progress bar at the bottom
+    with Progress(
+        SpinnerColumn(),
+        TextColumn("[bold blue]{task.description}"),
+        BarColumn(bar_width=40),
+        TextColumn("[progress.percentage]{task.percentage:>3.0f}%"),
+        console=console,
+        transient=False,
+    ) as progress:
+        task = progress.add_task("Waiting for server response...", total=100)
+
+        while status_val not in ["SUCCESS", "FAILED", "CANCELLED"]:
+            try:
+                # 1. Update Status
+                dep_status = client.deployments.get_status(deployment_id)
+                status_val = dep_status.get("status", "PENDING")
+                progress_val = dep_status.get("progress", 0)
+                state = dep_status.get("state", "preparing")
+
+                # Use a more descriptive description for the progress task
+                desc = f"Phase: {state}"
+                if status_val == "FAILED":
+                    desc = "[bold red]FAILED[/bold red]"
+                elif status_val == "SUCCESS":
+                    desc = "[bold green]SUCCESS[/bold green]"
+
+                progress.update(task, completed=progress_val, description=desc)
+
+                # 2. Update Logs
+                log_content = client.deployments.get_logs(deployment_id)
+                if log_content and len(log_content) > last_log_len:
+                    new_logs = log_content[last_log_len:].strip()
+                    for line in new_logs.split("\n"):
+                        # Process and colorize lines
+                        clean_line = line.strip()
+                        if not clean_line:
+                            continue
+
+                        if any(x in clean_line for x in ["ERROR", "FAILED", "✗"]):
+                            progress.console.print(f"[bold red]{clean_line}[/bold red]")
+                        elif any(x in clean_line for x in ["WARN", "WARNING", "⚠"]):
+                            progress.console.print(f"[yellow]{clean_line}[/yellow]")
+                        elif any(x in clean_line for x in ["SUCCESS", "COMPLETED", "✓", "passed!"]):
+                            progress.console.print(f"[bold green]{clean_line}[/bold green]")
+                        elif "PHASE" in clean_line:
+                            progress.console.print(f"\n[bold blue]{clean_line}[/bold blue]")
+                        elif "[InfraEngine]" in clean_line:
+                            progress.console.print(f"[dim]{clean_line}[/dim]")
+                        else:
+                            progress.console.print(f"[cyan]›[/cyan] {clean_line}")
+
+                    last_log_len = len(log_content)
+
+                if status_val in ["SUCCESS", "FAILED", "CANCELLED"]:
+                    break
+
+                time.sleep(1.5)  # Slightly faster polling for better feel
+            except Exception as e:
+                # progress.console.print(f"[dim]Transient connection issue: {e}[/dim]")
+                time.sleep(3)
+                continue
+
+    console.print()
+    if status_val == "SUCCESS":
+        console.print("[bold green]✨ SUCCESS: Your application is live![/bold green]")
+        # Try to get the IP address
+        try:
+            final_status = client.deployments.get_status(deployment_id)
+            ip = final_status.get("ip_address")
+            if ip:
+                console.print(f"[bold]Accessible at:[/bold] [link=http://{ip}]http://{ip}[/link]")
+        except:
+            pass
+    elif status_val == "FAILED":
+        console.print("\n[bold red]❌ FAILURE DETECTED: Entering AI Diagnosis Mode...[/bold red]")
+
+    return status_val
+
+
+def zen_nod_workflow(
+    logs: str, 
+    client: XenfraClient, 
+    attempt: int,
+    package_manager: str = None,
+    dependency_file: str = None,
+    services: list = None
+) -> bool:
+    """
+    Execute the Zen Nod auto-healing workflow.
+
+    Args:
+        logs: Deployment error logs
+        client: Authenticated SDK client
+        attempt: Current attempt number
+        package_manager: Project package manager
+        dependency_file: Project dependency file
+        services: List of services in the project (for multi-service context)
+
+    Returns:
+        True if patch was applied and user wants to retry, False otherwise
+    """
+    console.print()
+    console.print(f"[cyan]🤖 Analyzing failure (attempt {attempt}/{MAX_RETRY_ATTEMPTS})...[/cyan]")
+
+    # Slice logs to last 300 lines for focused diagnosis (Fix #26)
+    log_lines = logs.split("\n")
+    if len(log_lines) > 300:
+        logs = "\n".join(log_lines[-300:])
+        console.print("[dim]Note: Analyzing only the last 300 lines of logs for efficiency.[/dim]")
+
+    # Scrub sensitive data from logs
+    scrubbed_logs = scrub_logs(logs)
+
+    # Diagnose with AI
+    try:
+        diagnosis_result = client.intelligence.diagnose(
+            logs=scrubbed_logs,
+            package_manager=package_manager,
+            dependency_file=dependency_file,
+            services=services
+        )
+    except Exception as e:
+        console.print(f"[yellow]Could not diagnose failure: {e}[/yellow]")
+        return False
+
+    # Show diagnosis
+    show_diagnosis_panel(diagnosis_result.diagnosis, diagnosis_result.suggestion)
+
+    # Check if there's an automatic patch
+    if diagnosis_result.patch and diagnosis_result.patch.file:
+        show_patch_preview(diagnosis_result.patch.model_dump())
+
+        # Zen Nod confirmation
+        if click.confirm("Apply this fix and retry deployment?", default=True):
+            try:
+                # Apply patch (with automatic backup)
+                backup_path = apply_patch(diagnosis_result.patch.model_dump())
+                console.print("[bold green]✓ Patch applied[/bold green]")
+                if backup_path:
+                    console.print(f"[dim]Backup saved: {backup_path}[/dim]")
+                return True  # Signal to retry
+            except Exception as e:
+                console.print(f"[bold red]Failed to apply patch: {e}[/bold red]")
+                return False
+        else:
+            console.print()
+            console.print("[yellow]❌ Patch declined. Follow the manual steps above.[/yellow]")
+            return False
+    else:
+        console.print()
+        console.print(
+            "[yellow]No automatic fix available. Please follow the manual steps above.[/yellow]"
+        )
+        return False
+
+
+@click.command()
+@click.argument("project_id", type=int)
+@click.option("--show-details", is_flag=True, help="Show project details before confirmation")
+@click.confirmation_option(prompt="Are you sure you want to delete this project and its infrastructure?")
+def delete(project_id, show_details):
+    """
+    Delete a project and its infrastructure.
+
+    This command will:
+    - Destroy the DigitalOcean droplet
+    - Remove database records
+
+    WARNING: This action cannot be undone.
+    """
+    # Validate
+    is_valid, error_msg = validate_project_id(project_id)
+    if not is_valid:
+        console.print(f"[bold red]Invalid project ID: {error_msg}[/bold red]")
+        raise click.Abort()
+
+    try:
+        with get_client() as client:
+            # Optional: Show details
+            if show_details:
+                try:
+                    project = client.projects.show(project_id)
+
+                    # Display panel with project info
+                    details_table = Table(show_header=False, box=None)
+                    details_table.add_column("Property", style="cyan")
+                    details_table.add_column("Value")
+
+                    details_table.add_row("Project ID", str(project_id))
+                    if hasattr(project, 'name'):
+                        details_table.add_row("Name", project.name)
+                    if hasattr(project, 'droplet_id'):
+                        details_table.add_row("Droplet ID", str(project.droplet_id))
+                    if hasattr(project, 'ip_address'):
+                        details_table.add_row("IP Address", project.ip_address)
+                    if hasattr(project, 'created_at'):
+                        details_table.add_row("Created", str(project.created_at))
+
+                    panel = Panel(details_table, title="[bold]Project Details[/bold]", border_style="yellow")
+                    console.print(panel)
+                    console.print()
+                except XenfraAPIError as e:
+                    if e.status_code == 404:
+                        console.print(f"[yellow]Note: Project {project_id} not found in records.[/yellow]")
+                    else:
+                        console.print(f"[yellow]Warning: Could not fetch project details: {e.detail}[/yellow]")
+                    console.print()
+
+            # Delete
+            console.print(f"[cyan]Deleting project {project_id}...[/cyan]")
+            client.projects.delete(str(project_id))
+            console.print(f"[bold green]✓ Project {project_id} deleted successfully.[/bold green]")
+            console.print("[dim]The droplet has been destroyed and all records removed.[/dim]")
+
+    except XenfraAPIError as e:
+        if e.status_code == 404:
+            console.print(f"[yellow]Project {project_id} not found. It may have already been deleted.[/yellow]")
+        else:
+            console.print(f"[bold red]API Error: {e.detail}[/bold red]")
+            raise click.Abort()
+    except XenfraError as e:
+        console.print(f"[bold red]Error: {e}[/bold red]")
+        raise click.Abort()
+    except click.Abort:
+        console.print("[dim]Deletion cancelled.[/dim]")
+        raise
+    except Exception as e:
+        console.print(f"[bold red]Unexpected error: {e}[/bold red]")
+        raise click.Abort()
+
+
+@click.command()
+@click.option("--project-name", help="Project name (defaults to current directory name)")
+@click.option("--git-repo", help="Git repository URL (if deploying from git)")
+@click.option("--branch", default="main", help="Git branch (default: main)")
+@click.option("--framework", help="Framework override (fastapi, flask, django)")
+@click.option("--region", help="DigitalOcean region override")
+@click.option("--size", help="DigitalOcean size slug override")
+@click.option("--no-heal", is_flag=True, help="Disable auto-healing on failure")
+@click.option("--cleanup-on-failure", is_flag=True, help="Automatically cleanup resources if deployment fails")
+@click.option("--force-sandbox", is_flag=True, help="Force full sandbox validation (Tier 3)")
+@click.option("--skip-sandbox", is_flag=True, help="Skip sandbox validation for faster iteration")
+def deploy(project_name, git_repo, branch, framework, region, size, no_heal, cleanup_on_failure, force_sandbox, skip_sandbox):
+    """
+    Deploy current project to DigitalOcean with 3-tier dry-run validation.
+
+    Every deployment is validated before touching real infrastructure:
+
+    \b
+    [bold]Tier 1[/bold]: Local validation (~100ms) - Syntax, imports, config
+    [bold]Tier 2[/bold]: Railpack plan (~1-2s) - Build plan generation
+    [bold]Tier 3[/bold]: E2B Sandbox (~30-60s) - Full build validation (smart detection)
+
+    This "Dry Run First" architecture ensures deployments never fail due to
+    avoidable errors. Set E2B_API_KEY for full sandbox validation.
+
+    \b
+    Deploys your application with zero configuration. The CLI will:
+    1. Run 3-tier dry-run validation
+    2. Create a deployment
+    3. Auto-diagnose and fix failures (unless --no-heal is set)
+
+    Environment Variables:
+    - E2B_API_KEY: API key for sandbox validation
+    - XENFRA_AI=off: Disable AI features
+    - XENFRA_SANDBOX=force: Always run Tier 3
+    - XENFRA_SANDBOX=skip: Skip Tier 3 validation
+    - XENFRA_SANDBOX=auto: Smart detection (default)
+    """
+    # Check XENFRA_AI environment variable
+    ai_mode = os.environ.get("XENFRA_AI", "on").lower()
+    no_ai = ai_mode == "off"
+    if no_ai:
+        console.print("[yellow]XENFRA_AI=off is set. Auto-healing disabled.[/yellow]")
+        no_heal = True
+
+    # ==========================================
+    # 3-TIER DRY-RUN VALIDATION ("Dry Run First")
+    # ==========================================
+    project_path = Path(os.getcwd())
+
+    # Check for xenfra.yaml
+    if not has_xenfra_config():
+        console.print("[yellow]No xenfra.yaml found.[/yellow]")
+        if click.confirm("Run 'xenfra init' to create configuration?", default=True):
+            from .intelligence import init
+
+            ctx = click.get_current_context()
+            ctx.invoke(init, manual=no_ai, accept_all=False)
+        else:
+            console.print("[dim]Deployment cancelled.[/dim]")
+            return
+
+    # Load configuration from xenfra.yaml if it exists
+    config = {}
+    if has_xenfra_config():
+        try:
+            config = read_xenfra_yaml()
+        except Exception as e:
+            console.print(f"[yellow]Warning: Could not read xenfra.yaml: {e}[/dim]")
+
+    # Resolve values with precedence: 1. CLI Flag, 2. xenfra.yaml, 3. Default
+    project_name = project_name or config.get("name") or os.path.basename(os.getcwd())
+    framework = framework or config.get("framework")
+    # Track if is_dockerized was explicitly set in config (to avoid AI override)
+    is_dockerized_from_config = "is_dockerized" in config
+    is_dockerized = config.get("is_dockerized", True)
+    region = region or config.get("region") or "nyc3"
+    
+    # Resolve size slug (complex mapping)
+    if not size:
+        if config.get("size"):
+            size = config.get("size")
+        else:
+            instance_size = config.get("instance_size", "basic")
+            resources = config.get("resources", {})
+            cpu = resources.get("cpu", 1)
+            
+            if instance_size == "standard" or cpu >= 2:
+                size = "s-2vcpu-4gb"
+            elif instance_size == "premium" or cpu >= 4:
+                size = "s-4vcpu-8gb"
+            else:
+                size = "s-1vcpu-1gb"
+
+    # Extract port, command, database from config
+    # Track if port was explicitly set to avoid AI override
+    port_from_config = config.get("port")
+    port = port_from_config or 8000
+    command = config.get("command")  # Auto-detected if not provided
+    entrypoint = config.get("entrypoint")  # e.g., "todo.main:app"
+    database_config = config.get("database", {})
+    database = database_config.get("type") if isinstance(database_config, dict) else None
+    package_manager = config.get("package_manager", "pip")
+    dependency_file = config.get("dependency_file", "requirements.txt")
+    
+    # Microservices support: extract services and mode from xenfra.yaml
+    services = config.get("services")  # List of service definitions
+    mode = config.get("mode", "monolithic")  # monolithic, single-droplet, multi-droplet
+    
+    # If services are defined and > 1, this is a microservices deployment
+    if services and len(services) > 1:
+        console.print(f"\n[bold cyan]🔍 Detected microservices project ({len(services)} services)[/bold cyan]")
+        
+        # Display services table
+        from rich.table import Table
+        svc_table = Table(show_header=True, header_style="bold cyan", box=None)
+        svc_table.add_column("Service", style="white")
+        svc_table.add_column("Port", style="green")
+        svc_table.add_column("Framework", style="yellow")
+        
+        for svc in services:
+            svc_table.add_row(
+                svc.get("name", "?"),
+                str(svc.get("port", "?")),
+                svc.get("framework", "?")
+            )
+        
+        console.print(svc_table)
+        console.print()
+        
+        # If mode not explicitly set in config, ask user
+        if mode == "monolithic" or mode not in ["single-droplet", "multi-droplet"]:
+            console.print("[bold]Choose deployment mode:[/bold]")
+            console.print("  [cyan]1.[/cyan] Single Droplet - All services on one machine [dim](cost-effective)[/dim]")
+            console.print("  [cyan]2.[/cyan] Multi Droplet  - Each service on its own machine [dim](scalable)[/dim]")
+            console.print()
+            
+            mode_choice = Prompt.ask(
+                "Deployment mode",
+                choices=["1", "2"],
+                default="1"
+            )
+            
+            if mode_choice == "1":
+                mode = "single-droplet"
+                console.print("[green]✓ Using single-droplet mode[/green]\n")
+            else:
+                mode = "multi-droplet"
+                console.print("[green]✓ Using multi-droplet mode[/green]")
+                console.print(f"[dim]This will create {len(services)} separate droplets[/dim]\n")
+        else:
+            console.print(f"[dim]Using configured mode: {mode}[/dim]\n")
+
+    # Default project name to current directory
+    if not project_name:
+        project_name = os.path.basename(os.getcwd())
+
+    # Validate project name
+    is_valid, error_msg = validate_project_name(project_name)
+    if not is_valid:
+        console.print(f"[bold red]Invalid project name: {error_msg}[/bold red]")
+        raise click.Abort()
+
+    # Validate git repo if provided
+    if git_repo:
+        is_valid, error_msg = validate_git_repo_url(git_repo)
+        if not is_valid:
+            console.print(f"[bold red]Invalid git repository URL: {error_msg}[/bold red]")
+            raise click.Abort()
+        console.print(f"[cyan]Deploying {project_name} from git repository...[/cyan]")
+        console.print(f"[dim]Repository: {git_repo} (branch: {branch})[/dim]")
+    else:
+        # Note: Local folder deployment only works when engine runs locally
+        # In cloud API mode, this will fail with a clear error from the server
+        console.print(f"[cyan]Deploying {project_name}...[/cyan]")
+        console.print("[dim]Note: Git repository recommended for cloud deployments[/dim]")
+
+    # Validate branch name
+    is_valid, error_msg = validate_branch_name(branch)
+    if not is_valid:
+        console.print(f"[bold red]Invalid branch name: {error_msg}[/bold red]")
+        raise click.Abort()
+
+    # Validate framework if provided
+    if framework:
+        is_valid, error_msg = validate_framework(framework)
+        if not is_valid:
+            console.print(f"[bold red]Invalid framework: {error_msg}[/bold red]")
+            raise click.Abort()
+
+    # Retry loop for auto-healing
+    attempt = 0
+    deployment_id = None
+    
+    # Context manager for temp dir (if cloning)
+    import tempfile
+    import shutil
+    import subprocess
+    
+    temp_dir_obj = None
+    original_project_path = project_path
+    
+    try:
+        with get_client() as client:
+            from ..utils.file_sync import scan_project_files_cached, ensure_gitignore_ignored
+            
+            # --- 1. PREPARE SOURCE CODE ---
+            if git_repo:
+                console.print(f"[cyan]📥 Cloning {git_repo}...[/cyan]")
+                temp_dir_obj = tempfile.TemporaryDirectory()
+                temp_dir_path = Path(temp_dir_obj.name)
+                
+                try:
+                    # Shallow clone for speed
+                    cmd = ["git", "clone", "--depth", "1"]
+                    if branch:
+                        cmd.extend(["-b", branch])
+                    cmd.extend([git_repo, str(temp_dir_path)])
+                    
+                    subprocess.run(cmd, check=True, capture_output=True)
+                    project_path = temp_dir_path
+                    console.print("[green]✓ Repository cloned[/green]")
+                except subprocess.CalledProcessError as e:
+                    console.print(f"[bold red]Failed to clone repository: {e}[/bold red]")
+                    raise click.Abort()
+            else:
+                # Local directory
+                 # Protect privacy: ensure .xenfra is in .gitignore
+                if ensure_gitignore_ignored():
+                    console.print("[dim]   - Added .xenfra to .gitignore for privacy[/dim]")
+
+            # --- 2. SCAN FILES ---
+            console.print("[cyan]📁 Scanning project files...[/cyan]")
+            
+            # We need to temporarily change CWD for scan_project_files_cached to work if we cloned
+            # OR pass path to it. Assuming it scans CWD, we might need chdir
+            import os
+            cwd = os.getcwd()
+            try:
+                if project_path != Path(cwd):
+                    os.chdir(project_path)
+                file_manifest = scan_project_files_cached()
+            finally:
+                os.chdir(cwd)
+
+            console.print(f"[dim]Found {len(file_manifest)} files[/dim]")
+            
+            if not file_manifest:
+                console.print("[bold red]Error: No files found.[/bold red]")
+                raise click.Abort()
+
+            # --- 3. UPLOAD FILES (Required for Tier 3 & Deployment) ---
+            # Check which files need uploading
+            console.print("[cyan]🔍 Checking file cache...[/cyan]")
+            check_result = client.files.check(file_manifest)
+            missing = check_result.get('missing', [])
+            cached = check_result.get('cached', 0)
+            
+            if cached > 0:
+                console.print(f"[green]✓ {cached} files already cached on server[/green]")
+            
+            # Upload missing files
+            if missing:
+                console.print(f"[cyan]☁️ Uploading {len(missing)} files...[/cyan]")
+                # We need absolute paths for upload reading
+                # If we cloned, file_manifest paths are relative to temp_dir
+                # But upload_files needs to know where to read from.
+                # Let's patch file_manifest with abs_paths for the uploader
+                
+                upload_manifest = []
+                for f in file_manifest:
+                    f_copy = f.copy()
+                    f_copy["abs_path"] = str(project_path / f["path"])
+                    upload_manifest.append(f_copy)
+
+                uploaded = client.files.upload_files(
+                    upload_manifest,
+                    missing,
+                    progress_callback=lambda done, total: console.print(f"[dim]   Progress: {done}/{total}[/dim]") if done % 10 == 0 or done == total else None
+                )
+                console.print(f"[green]✓ Uploaded {uploaded} files[/green]")
+            else:
+                console.print("[green]✓ All files synced[/green]")
+            
+            # --- 4. RUN DRY RUN ---
+            console.print(Panel(
+                "[bold]🔍 Running 3-Tier Dry-Run Validation[/bold]\n"
+                "Tier 1: Local syntax check (~100ms)\n"
+                "Tier 2: Railpack build plan (~1-2s)\n"
+                "Tier 3: E2B sandbox via Xenfra API (smart detection)",
+                title="[bold cyan]Dry Run First Architecture[/bold cyan]",
+                border_style="cyan"
+            ))
+            
+            with console.status("[cyan]Running validation...[/cyan]"):
+                dry_run_result = run_dry_run(
+                    project_path,
+                    force_sandbox=force_sandbox, # or (force_sandbox or check_result['missing'])?
+                    skip_sandbox=skip_sandbox,
+                    api_client=client,
+                    file_manifest=file_manifest
+                )
+            
+            # Display results
+            console.print(dry_run_result.get_summary())
+            console.print()
+            
+            # Check if dry-run failed
+            if not dry_run_result.success:
+                console.print("[bold red]❌ Dry-run validation failed[/bold red]")
+                console.print("[yellow]Fix the errors above before deploying.[/yellow]")
+                console.print()
+                console.print("[dim]Tip: Run with --skip-sandbox to skip Tier 3[/dim]")
+                raise click.Abort()
+            
+            console.print("[bold green]✅ Dry-run validation passed[/bold green]")
+            console.print()
+            
+            # Prepare manifest for API (remove local-only fields if any, strictly path/sha/size)
+            api_file_manifest = [{"path": f["path"], "sha": f["sha"], "size": f["size"]} for f in file_manifest]
+
+            while attempt < MAX_RETRY_ATTEMPTS:
+                attempt += 1
+
+                if attempt > 1:
+                    console.print(
+                        f"\n[cyan]🔄 Retrying deployment (attempt {attempt}/{MAX_RETRY_ATTEMPTS})...[/cyan]"
+                    )
+                else:
+                    console.print("[cyan]Creating deployment...[/cyan]")
+
+                # Detect framework if not provided (AI-powered Zen Mode)
+                if not framework:
+                    if no_ai:
+                        # ... static logic existing ...
+                        console.print("[yellow]⚠ AI disabled. Using static detection...[/yellow]")
+                        # Simple static checks based on Tier 2 result potentially?
+                        if dry_run_result.tier2:
+                             framework = dry_run_result.tier2.framework
+                             port = dry_run_result.tier2.port
+                             is_dockerized = True
+                             console.print(f"[green]✓ Detected {framework.upper()} (from Dry Run)[/green]")
+                        else:
+                             framework = "fastapi"
+                             port = 8000
+                    else:
+                        # AI-powered detection
+                        # We can use Tier 2 result for speed instead of full AI scan if available
+                        if dry_run_result.tier2 and dry_run_result.tier2.framework != "unknown":
+                            framework = dry_run_result.tier2.framework
+                            # config overrides...
+                            if not is_dockerized_from_config:
+                                is_dockerized = True # Railpack is docker based
+                            if not port_from_config:
+                                port = dry_run_result.tier2.port
+                            console.print(f"[green]✓ Detected {framework.upper()} (from Railpack)[/green]")
+                        else:
+                             # Fallback to AI
+                             console.print("[cyan]🔍 AI Auto-detecting project type...[/cyan]")
+                             try:
+                                 # We need to scan code snippet from PROJECT_PATH (temp or local)
+                                 from ..utils.codebase import scan_codebase
+                                 # scan_codebase needs CWD to be correct
+                                 curr = os.getcwd()
+                                 os.chdir(project_path)
+                                 try:
+                                     code_snippets = scan_codebase()
+                                 finally:
+                                     os.chdir(curr)
+                                     
+                                 if code_snippets:
+                                     analysis = client.intelligence.analyze_codebase(code_snippets)
+                                     framework = analysis.framework
+                                     if not is_dockerized_from_config: is_dockerized = analysis.is_dockerized
+                                     if not port_from_config and analysis.port: port = analysis.port
+                                     if not size and analysis.instance_size: size = "s-1vcpu-1gb" if analysis.instance_size == "basic" else "s-2vcpu-4gb"
+                                     mode_str = "Docker" if is_dockerized else "Bare Metal"
+                                     console.print(f"[green]✓ Detected {framework.upper()} project ({mode_str} Mode)[/green] (Port: {port})")
+                                 else:
+                                     framework = "fastapi"
+                                     is_dockerized = True
+                             except Exception as e:
+                                 framework = "fastapi"
+                                 is_dockerized = True
+
+                # Delta upload was already done above!
+                # file_manifest is ready (api_file_manifest)
+
+                # Create deployment with real-time streaming
+                try:
+                    status_result, deployment_id, logs_data = _stream_deployment(
+                        client=client,
+                        project_name=project_name,
+                        git_repo=git_repo,
+                        branch=branch,
+                        framework=framework,
+                        region=region,
+                        size_slug=size,
+                        is_dockerized=is_dockerized,
+                        port=port,
+                        command=command,
+                        entrypoint=entrypoint,  # Pass entrypoint to deployment API
+                        database=database,
+                        package_manager=package_manager,
+                        dependency_file=dependency_file,
+                        file_manifest=api_file_manifest,
+                        cleanup_on_failure=cleanup_on_failure,
+                        services=services,  # Microservices support
+                        mode=mode,  # Deployment mode
+                    )
+
+
+                    if status_result == "FAILED" and not no_heal:
+                        # Hand off to the Zen Nod AI Agent
+                        should_retry = zen_nod_workflow(
+                            logs_data, 
+                            client, 
+                            attempt,
+                            package_manager=package_manager,
+                            dependency_file=dependency_file,
+                            services=services
+                        )
+
+                        if should_retry:
+                            # The agent applied a fix, loop back for attempt + 1
+                            continue
+                        else:
+                            # Agent couldn't fix it or user declined
+                            raise click.Abort()
+
+                    # If we got here with success, break the retry loop
+                    if status_result == "SUCCESS":
+                        break
+                    else:
+                        raise click.Abort()
+
+                except XenfraAPIError as e:
+                    # Deployment failed - try to provide helpful error
+                    from ..utils.errors import detect_error_type, show_error_with_solution
+                    
+                    console.print(f"[bold red]✗ Deployment failed[/bold red]")
+                    
+                    # Try to detect error type and show solution
+                    error_type, error_kwargs = detect_error_type(str(e.detail))
+                    if error_type:
+                        show_error_with_solution(error_type, **error_kwargs)
+                    else:
+                        console.print(f"[red]{e.detail}[/red]")
+
+                    # Check if we should auto-heal
+                    if no_heal or attempt >= MAX_RETRY_ATTEMPTS:
+                        # No auto-healing or max retries reached
+                        if attempt >= MAX_RETRY_ATTEMPTS:
+                            console.print(
+                                f"\n[bold red]❌ Maximum retry attempts ({MAX_RETRY_ATTEMPTS}) reached.[/bold red]"
+                            )
+                            console.print(
+                                "[yellow]Unable to auto-fix the issue. Please review the errors above.[/yellow]"
+                            )
+                        raise
+                    else:
+                        # Try to get logs for diagnosis
+                        error_logs = str(e.detail)
+                        try:
+                            if deployment_id:
+                                # This should be a method in the SDK that returns a string
+                                logs_response = client.deployments.get_logs(deployment_id)
+                                if isinstance(logs_response, dict):
+                                    error_logs = logs_response.get("logs", str(e.detail))
+                                else:
+                                    error_logs = str(logs_response)  # Assuming it can be a string
+                        except Exception as log_err:
+                            console.print(
+                                f"[yellow]Warning: Could not fetch detailed logs for diagnosis: {log_err}[/yellow]"
+                            )
+                            # Fallback to the initial error detail
+                            pass
+
+                        # Run Zen Nod workflow
+                        should_retry = zen_nod_workflow(
+                            error_logs, 
+                            client, 
+                            attempt,
+                            package_manager=package_manager,
+                            dependency_file=dependency_file,
+                            services=services
+                        )
+
+                        if not should_retry:
+                            # User declined patch or no patch available
+                            console.print("\n[dim]Deployment cancelled.[/dim]")
+                            raise click.Abort()
+
+                        # Continue to next iteration (retry)
+                        continue
+
+    except XenfraAPIError as e:
+        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
+    except XenfraError as e:
+        console.print(f"[bold red]Error: {e}[/bold red]")
+    except click.Abort:
+        pass
+    except Exception as e:
+        console.print(f"[bold red]Unexpected error: {e}[/bold red]")
+
+
+@click.command()
+@click.argument("deployment-id")
+@click.option("--follow", "-f", is_flag=True, help="Follow log output (stream)")
+@click.option("--tail", type=int, help="Show last N lines")
+def logs(deployment_id, follow, tail):
+    # Validate deployment ID
+    is_valid, error_msg = validate_deployment_id(deployment_id)
+    if not is_valid:
+        console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
+        raise click.Abort()
+    """
+    Stream deployment logs.
+
+    Shows logs for a specific deployment. Use --follow to stream logs in real-time.
+    """
+    try:
+        with get_client() as client:
+            console.print(f"[cyan]Fetching logs for deployment {deployment_id}...[/cyan]")
+
+            log_content = client.deployments.get_logs(deployment_id)
+
+            if not log_content:
+                console.print("[yellow]No logs available yet.[/yellow]")
+                console.print("[dim]The deployment may still be starting up.[/dim]")
+                return
+
+            # Process logs
+            log_lines = log_content.strip().split("\n")
+
+            # Apply tail if specified
+            if tail:
+                log_lines = log_lines[-tail:]
+
+            # Display logs with syntax highlighting
+            console.print(f"\n[bold]Logs for deployment {deployment_id}:[/bold]\n")
+
+            if follow:
+                _follow_deployment(client, deployment_id)
+                return
+
+            # Display logs
+            for line in log_lines:
+                # Color-code based on log level
+                if "ERROR" in line or "FAILED" in line:
+                    console.print(f"[red]{line}[/red]")
+                elif "WARN" in line or "WARNING" in line:
+                    console.print(f"[yellow]{line}[/yellow]")
+                elif "SUCCESS" in line or "COMPLETED" in line:
+                    console.print(f"[green]{line}[/green]")
+                elif "INFO" in line:
+                    console.print(f"[cyan]{line}[/cyan]")
+                else:
+                    console.print(line)
+
+    except XenfraAPIError as e:
+        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
+    except XenfraError as e:
+        console.print(f"[bold red]Error: {e}[/bold red]")
+    except click.Abort:
+        pass
+
+
+@click.command()
+@click.argument("deployment-id", required=False)
+@click.option("--watch", "-w", is_flag=True, help="Watch status updates")
+def status(deployment_id, watch):
+    """
+    Show deployment status.
+
+    Displays current status, progress, and details for a deployment.
+    Use --watch to monitor status in real-time.
+    """
+    try:
+        if not deployment_id:
+            console.print("[yellow]No deployment ID provided.[/yellow]")
+            console.print("[dim]Usage: xenfra status <deployment-id>[/dim]")
+            return
+
+        # Validate deployment ID
+        is_valid, error_msg = validate_deployment_id(deployment_id)
+        if not is_valid:
+            console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
+            raise click.Abort()
+
+        with get_client() as client:
+            console.print(f"[cyan]Fetching status for deployment {deployment_id}...[/cyan]")
+
+            deployment_status = client.deployments.get_status(deployment_id)
+
+            if watch:
+                _follow_deployment(client, deployment_id)
+                return
+
+            # Display status
+            status_value = deployment_status.get("status", "UNKNOWN")
+            state = deployment_status.get("state", "unknown")
+            progress = deployment_status.get("progress", 0)
+
+            # Status panel
+            status_color = {
+                "PENDING": "yellow",
+                "IN_PROGRESS": "cyan",
+                "SUCCESS": "green",
+                "FAILED": "red",
+                "CANCELLED": "dim",
+            }.get(status_value, "white")
+
+            # Create status table
+            table = Table(show_header=False, box=None)
+            table.add_column("Property", style="cyan")
+            table.add_column("Value")
+
+            table.add_row("Deployment ID", str(deployment_id))
+            table.add_row("Status", f"[{status_color}]{status_value}[/{status_color}]")
+            table.add_row("State", state)
+
+            if progress > 0:
+                table.add_row("Progress", f"{progress}%")
+
+            if "project_name" in deployment_status:
+                table.add_row("Project", deployment_status["project_name"])
+
+            if "created_at" in deployment_status:
+                table.add_row("Created", deployment_status["created_at"])
+
+            if "finished_at" in deployment_status:
+                table.add_row("Finished", deployment_status["finished_at"])
+
+            if "url" in deployment_status:
+                table.add_row("URL", f"[link]{deployment_status['url']}[/link]")
+
+            if "ip_address" in deployment_status:
+                table.add_row("IP Address", deployment_status["ip_address"])
+
+            panel = Panel(table, title="[bold]Deployment Status[/bold]", border_style=status_color)
+            console.print(panel)
+
+            # Show error if failed
+            if status_value == "FAILED" and "error" in deployment_status:
+                error_panel = Panel(
+                    deployment_status["error"],
+                    title="[bold red]Error[/bold red]",
+                    border_style="red",
+                )
+                console.print("\n", error_panel)
+
+                console.print("\n[bold]Troubleshooting:[/bold]")
+                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
+                console.print(f"  • Diagnose: [cyan]xenfra diagnose {deployment_id}[/cyan]")
+
+            # Show next steps based on status
+            elif status_value == "SUCCESS":
+                console.print("\n[bold green]Deployment successful! 🎉[/bold green]")
+                if "url" in deployment_status:
+                    console.print(f"  • Visit: [link]{deployment_status['url']}[/link]")
+
+            elif status_value in ["PENDING", "IN_PROGRESS"]:
+                console.print("\n[bold]Deployment in progress...[/bold]")
+                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
+                console.print(f"  • Check again: [cyan]xenfra status {deployment_id}[/cyan]")
+
+    except XenfraAPIError as e:
+        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
+    except XenfraError as e:
+        console.print(f"[bold red]Error: {e}[/bold red]")
+    except click.Abort:
+        pass
+
+
+@click.command()
+@click.argument("deployment-id")
+@click.option("--format", "output_format", type=click.Choice(["detailed", "summary"], case_sensitive=False), default="detailed", help="Report format (detailed or summary)")
+def report(deployment_id, output_format):
+    """
+    Generate deployment report with self-healing events.
+    
+    Shows comprehensive deployment information including:
+    - Deployment status and timeline
+    - Self-healing attempts and outcomes
+    - Patches applied during healing
+    - Statistics and metrics
+    """
+    try:
+        # Validate deployment ID
+        is_valid, error_msg = validate_deployment_id(deployment_id)
+        if not is_valid:
+            console.print(f"[bold red]Invalid deployment ID: {error_msg}[/bold red]")
+            raise click.Abort()
+
+        with get_client() as client:
+            console.print(f"[cyan]Generating report for deployment {deployment_id}...[/cyan]\n")
+
+            # Get deployment status
+            try:
+                deployment_status = client.deployments.get_status(deployment_id)
+            except XenfraAPIError as e:
+                console.print(f"[bold red]Error fetching deployment status: {e.detail}[/bold red]")
+                raise click.Abort()
+
+            # Get deployment logs
+            try:
+                logs = client.deployments.get_logs(deployment_id)
+            except XenfraAPIError:
+                logs = None
+
+            # Parse status
+            status_value = deployment_status.get("status", "UNKNOWN")
+            state = deployment_status.get("state", "unknown")
+            progress = deployment_status.get("progress", 0)
+
+            # Status color mapping
+            status_color = {
+                "PENDING": "yellow",
+                "IN_PROGRESS": "cyan",
+                "SUCCESS": "green",
+                "FAILED": "red",
+                "CANCELLED": "dim",
+            }.get(status_value, "white")
+
+            # Calculate statistics from logs
+            heal_attempts = logs.count("🤖 Analyzing failure") if logs else 0
+            patches_applied = logs.count("✓ Patch applied") if logs else 0
+            diagnoses = logs.count("🔍 Diagnosis") if logs else 0
+
+            # Create main report table
+            report_table = Table(show_header=True, box=None)
+            report_table.add_column("Property", style="cyan", width=25)
+            report_table.add_column("Value", style="white")
+
+            report_table.add_row("Deployment ID", str(deployment_id))
+            report_table.add_row("Status", f"[{status_color}]{status_value}[/{status_color}]")
+            report_table.add_row("State", state)
+            
+            if progress > 0:
+                report_table.add_row("Progress", f"{progress}%")
+
+            if "project_name" in deployment_status:
+                report_table.add_row("Project", deployment_status["project_name"])
+
+            if "created_at" in deployment_status:
+                report_table.add_row("Created", deployment_status["created_at"])
+
+            if "finished_at" in deployment_status:
+                report_table.add_row("Finished", deployment_status["finished_at"])
+
+            if "url" in deployment_status:
+                report_table.add_row("URL", f"[link]{deployment_status['url']}[/link]")
+
+            if "ip_address" in deployment_status:
+                report_table.add_row("IP Address", deployment_status["ip_address"])
+
+            # Self-healing statistics
+            report_table.add_row("", "")  # Separator
+            report_table.add_row("[bold]Self-Healing Stats[/bold]", "")
+            report_table.add_row("Healing Attempts", str(heal_attempts))
+            report_table.add_row("Patches Applied", str(patches_applied))
+            report_table.add_row("Diagnoses Performed", str(diagnoses))
+            
+            if heal_attempts > 0:
+                success_rate = (patches_applied / heal_attempts * 100) if heal_attempts > 0 else 0
+                report_table.add_row("Healing Success Rate", f"{success_rate:.1f}%")
+
+            # Display main report
+            console.print(Panel(report_table, title="[bold]Deployment Report[/bold]", border_style=status_color))
+
+            # Detailed format includes timeline and healing events
+            if output_format == "detailed" and logs:
+                console.print("\n[bold]Self-Healing Timeline[/bold]\n")
+                
+                # Extract healing events from logs
+                log_lines = logs.split("\n")
+                timeline_entries = []
+                
+                for i, line in enumerate(log_lines):
+                    if "🤖 Analyzing failure" in line:
+                        attempt_match = None
+                        # Try to find attempt number in surrounding lines
+                        for j in range(max(0, i-5), min(len(log_lines), i+10)):
+                            if "attempt" in log_lines[j].lower():
+                                timeline_entries.append(("Healing Attempt", log_lines[j].strip()))
+                                break
+                    elif "🔍 Diagnosis" in line or "Diagnosis" in line:
+                        # Extract diagnosis text from next few lines
+                        diagnosis_text = line.strip()
+                        if i+1 < len(log_lines) and log_lines[i+1].strip():
+                            diagnosis_text += "\n  " + log_lines[i+1].strip()[:100]
+                        timeline_entries.append(("Diagnosis", diagnosis_text))
+                    elif "✓ Patch applied" in line or "Patch applied" in line:
+                        timeline_entries.append(("Patch Applied", line.strip()))
+                    elif "🔄 Retrying deployment" in line:
+                        timeline_entries.append(("Retry", line.strip()))
+
+                if timeline_entries:
+                    timeline_table = Table(show_header=True, box=None)
+                    timeline_table.add_column("Event", style="cyan", width=20)
+                    timeline_table.add_column("Details", style="white")
+
+                    for event_type, details in timeline_entries[:20]:  # Limit to 20 entries
+                        timeline_table.add_row(event_type, details)
+
+                    console.print(timeline_table)
+                else:
+                    console.print("[dim]No self-healing events detected in logs.[/dim]")
+
+            # Show error if failed
+            if status_value == "FAILED":
+                console.print("\n[bold red]⚠ Deployment Failed[/bold red]")
+                if "error" in deployment_status:
+                    error_panel = Panel(
+                        deployment_status["error"],
+                        title="[bold red]Error Details[/bold red]",
+                        border_style="red",
+                    )
+                    console.print("\n", error_panel)
+
+                console.print("\n[bold]Troubleshooting:[/bold]")
+                console.print(f"  • View logs: [cyan]xenfra logs {deployment_id}[/cyan]")
+                console.print(f"  • Diagnose: [cyan]xenfra diagnose {deployment_id}[/cyan]")
+                console.print(f"  • View status: [cyan]xenfra status {deployment_id}[/cyan]")
+
+            # Success summary
+            elif status_value == "SUCCESS":
+                console.print("\n[bold green]✓ Deployment Successful[/bold green]")
+                if heal_attempts > 0:
+                    console.print(f"[dim]Deployment succeeded after {heal_attempts} self-healing attempt(s).[/dim]")
+                if "url" in deployment_status:
+                    console.print(f"  • Visit: [link]{deployment_status['url']}[/link]")
+
+            # Export summary format (JSON-like structure for programmatic use)
+            if output_format == "summary":
+                console.print("\n[bold]Summary Format:[/bold]")
+                import json
+                summary = {
+                    "deployment_id": deployment_id,
+                    "status": status_value,
+                    "healing_attempts": heal_attempts,
+                    "patches_applied": patches_applied,
+                    "success": status_value == "SUCCESS",
+                }
+                console.print(f"[dim]{json.dumps(summary, indent=2)}[/dim]")
+
+    except XenfraAPIError as e:
+        console.print(f"[bold red]API Error: {e.detail}[/bold red]")
+    except XenfraError as e:
+        console.print(f"[bold red]Error: {e}[/bold red]")
+    except click.Abort:
+        pass
+    except Exception as e:
+        console.print(f"[bold red]Unexpected error: {e}[/bold red]")
+
+
+
+@click.command(name="dry-run")
+@click.option("--force-sandbox", is_flag=True, help="Force full sandbox validation (Tier 3)")
+@click.option("--skip-sandbox", is_flag=True, help="Skip sandbox validation for faster iteration")
+@click.option("--verbose", "-v", is_flag=True, help="Show detailed output")
+def dry_run_command(force_sandbox, skip_sandbox, verbose):
+    """
+    Run 3-tier dry-run validation without deploying.
+
+    Validates your project through all three tiers without
+    actually deploying to infrastructure:
+
+    \b
+    [bold]Tier 1[/bold]: Local validation - Syntax, imports, config files
+    [bold]Tier 2[/bold]: Railpack plan - Generate build plan
+    [bold]Tier 3[/bold]: E2B Sandbox - Full build validation via Xenfra API
+
+    This is useful for:
+    - Validating changes before deploying
+    - CI/CD pipelines
+    - Checking for syntax errors quickly
+
+    Tier 3 runs on Xenfra's servers - no local setup needed.
+    """
+    from pathlib import Path
+    from ..blueprints import run_dry_run
+    from ..utils.file_sync import scan_project_files_cached, ensure_gitignore_ignored
+    
+    project_path = Path(os.getcwd())
+    
+    # Get authenticated client for Tier 3 sandbox
+    try:
+        client = get_client()
+    except click.Abort:
+        return
+    
+    # Scan project files
+    if ensure_gitignore_ignored():
+        console.print("[dim]   - Added .xenfra to .gitignore for privacy[/dim]")
+    
+    console.print("[cyan]📁 Scanning project files...[/cyan]")
+    file_manifest = scan_project_files_cached()
+    console.print(f"[dim]Found {len(file_manifest)} files[/dim]")
+    
+    if not file_manifest:
+        console.print("[bold red]Error: No files found to validate.[/bold red]")
+        raise click.Abort()
+    
+    console.print(Panel(
+        "[bold]🔍 Running 3-Tier Dry-Run Validation[/bold]\n"
+        "Validating your project without deploying...",
+        title="[bold cyan]Dry Run[/bold cyan]",
+        border_style="cyan"
+    ))
+    
+    # Run dry-run
+    with console.status("[cyan]Running validation...[/cyan]"):
+        result = run_dry_run(
+            project_path,
+            force_sandbox=force_sandbox,
+            skip_sandbox=skip_sandbox,
+            api_client=client,
+            file_manifest=file_manifest
+        )
+    
+    # Display results
+    console.print(result.get_summary())
+    console.print()
+    
+    # Detailed output if verbose
+    if verbose and result.tier2:
+        console.print("[bold]Build Plan Details:[/bold]")
+        console.print(f"  Framework: {result.tier2.framework}")
+        console.print(f"  Language: {result.tier2.language}")
+        console.print(f"  Package Manager: {result.tier2.package_manager}")
+        console.print(f"  Port: {result.tier2.port}")
+        if result.tier2.start_command:
+            console.print(f"  Start Command: {result.tier2.start_command}")
+        console.print()
+    
+    # Final status
+    if result.success:
+        console.print("[bold green]✅ All validations passed - Ready to deploy![/bold green]")
+        console.print("[dim]Run 'xenfra deploy' to deploy to infrastructure.[/dim]")
+    else:
+        console.print("[bold red]❌ Validation failed - Fix errors before deploying[/bold red]")
+        raise click.Abort()