## ADDED Requirements

### Requirement: Server creation follows a staged deployment workflow
The management console SHALL collect a server definition in the order target, deployment method, mode-relevant configuration, and review. The console MUST preserve entered values while the operator moves between workflow steps and MUST prevent submission until the selected mode's required values are present.

#### Scenario: Create a guided server
- **WHEN** an operator selects a game plugin and compatible Run node, chooses guided installation, completes the plugin configuration, and confirms review
- **THEN** the console creates the server through the existing lifecycle workflow using the selected target and deployment definition

### Requirement: Draft creation is explicit
The management console SHALL present saving without a Run node as an explicit draft choice and MUST explain that no deployment job will be dispatched.

#### Scenario: Save an unbound definition
- **WHEN** an operator selects the draft choice during target selection and confirms the review
- **THEN** the console saves a draft and labels the final action as saving rather than deploying

### Requirement: Deployment configuration is mode-specific
The management console SHALL show only the deployment inputs needed by the selected deployment method. It MUST use server directory as the primary directory input and MUST expose execution working directory only in advanced custom-command configuration, with copy that it inherits the server directory when omitted.

#### Scenario: Configure an existing server
- **WHEN** an operator chooses existing-server deployment
- **THEN** the console requests the existing server directory and does not render installation or custom command inputs

#### Scenario: Configure a new installation
- **WHEN** an operator chooses guided-install deployment
- **THEN** the console labels the optional target as an installation directory, renders the plugin's game configuration fields, and does not present it as an existing-server adoption path

#### Scenario: Configure a custom launcher
- **WHEN** an operator chooses custom-command deployment
- **THEN** the console renders the required startup command and optional advanced execution directory, shell, install, stop, and status inputs

### Requirement: Existing server deployment editing is discoverable and safe
The management console SHALL expose an edit-deployment action from a server card and server-detail header. It MUST reuse the staged deployment workflow, use only safe deployment projections for prefill, and MUST prevent opening an editable deployment workflow for active servers.

#### Scenario: Edit a stopped server
- **WHEN** an operator invokes edit deployment for a stopped server
- **THEN** the console opens the staged workflow with its selected node, mode, create inputs, and protected-input configured state without revealing protected values

#### Scenario: Attempt to edit a running server
- **WHEN** an operator views a running server
- **THEN** the console identifies that deployment settings require stopping the server before editing and does not submit a deployment mutation

### Requirement: Review preserves deployment-input protection
The workflow review SHALL identify the chosen plugin, endpoint or draft status, deployment method, game configuration, and configured protected input state. It MUST NOT render stored or newly entered full host paths, command text, credentials, or runtime-binding secrets.

#### Scenario: Review custom deployment
- **WHEN** an operator reaches review with a custom-command deployment
- **THEN** the review reports that server directory, startup command, and optional execution fields are configured or replaced without exposing their values