Salesforce开发服务器工具包

# 1. Clone the setup repository
git clone https://github.com/franturino/mcp-server-setup.git
cd mcp-server-setup

# 2. Edit setup.ps1 - Update these variables:
#    - $BaseStoragePath: Your VS Code Server storage path
#    - $DxProjectPath: Your Salesforce dx-project path

# 3. Open PowerShell as Administrator and run:
Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser
.\setup.ps1

# 4. Wait for completion and check the log file
# Success! The MCP server is now installed and configured.

# 1. Clone the setup repository
git clone https://github.com/franturino/mcp-server-setup.git
cd mcp-server-setup

# 2. Edit setup.sh - Update these variables:
#    - BASE_STORAGE_PATH: Your Code Server storage path
#    - DX_PROJECT_PATH (if needed): Your Salesforce dx-project path

# 3. Make scripts executable
chmod +x setup.sh update.sh

# 4. Run setup
./setup.sh

# 5. Wait for completion and check the log file
# Success! The MCP server is now installed and configured.

git clone https://github.com/franturino/mcp-server-setup.git

cd mcp-server-setup

$BaseStoragePath = "C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt"
$DxProjectPath = "C:\Users\codebuilder\dx-project"

# 1. Open PowerShell as Administrator (required for scheduler permissions)
#    Right-click PowerShell → Run as Administrator

# 2. Enable script execution (if not already enabled)
Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

# 3. Navigate to the setup directory
cd C:\path\to\mcp-server-setup

# 4. Run the installation
.\setup.ps1

# 5. Monitor the output - wait for "INSTALLATION SCRIPT COMPLETED"

$BaseStoragePath/
├── MCP/
│   └── mcp_server_toolkit/          ← Cloned repository
│       ├── build/                   ← Compiled JavaScript (read-only)
│       ├── src/                     ← TypeScript source (read-only)
│       ├── package.json
│       └── ...other files
└── settings/
    └── a4d_mcp_settings.json        ← Updated with server config

BASE_STORAGE_PATH="/home/codebuilder/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt"

# 1. Navigate to the setup directory
cd /path/to/mcp-server-setup

# 2. Make scripts executable
chmod +x setup.sh update.sh

# 3. Run the installation
./setup.sh

# 4. Monitor the output - wait for "INSTALLATION SCRIPT COMPLETED"

$BASE_STORAGE_PATH/
├── MCP/
│   └── mcp_server_toolkit/          ← Cloned repository
│       ├── build/                   ← Compiled JavaScript
│       ├── src/                     ← TypeScript source
│       ├── package.json
│       └── ...other files
└── settings/
    └── a4d_mcp_settings.json        ← Updated with server config

1. Verifies installation exists at $REPO_PATH
2. Checks if mcp_server_toolkit is configured in a4d_mcp_settings.json
3. Exits with error if installation not found (must run setup.ps1/sh first)

1. Restores write permissions to build/ and src/ directories
2. Pulls latest changes from GitHub (git pull)
3. Installs/updates npm dependencies (npm install)
4. Rebuilds the project (npm run build)
5. Resets directories to read-only for protection
6. Creates timestamped log file with all details

# Open PowerShell and run:
cd C:\path\to\mcp-server-setup
.\update.ps1

# Check the log file for completion:
# update_YYYYMMDD_HHMMSS.log

# Navigate and run:
cd /path/to/mcp-server-setup
./update.sh

# Check the log file for completion:
# update_YYYYMMDD_HHmmss.log

# Navigate to your MCP directory
$McpPath = "C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt\MCP"
New-Item -Path $McpPath -ItemType Directory -Force | Out-Null
Set-Location $McpPath

# Clone the repository
git clone https://github.com/franturino/mcp_server_toolkit.git
cd mcp_server_toolkit

npm install

npm run build

# Copy workflow files to your dx-project
Copy-Item -Path "workflows\*" -Destination "C:\Users\codebuilder\dx-project\.a4drules\workflows\" -Recurse -Force

C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt\settings\a4d_mcp_settings.json

{
    "mcpServers": {
        "mcp_server_toolkit": {
            "disabled": false,
            "timeout": 600,
            "type": "stdio",
            "command": "node",
            "args": [
                "C:\\Users\\codebuilder\\.vscode-server\\data\\User\\globalStorage\\salesforce.salesforcedx-einstein-gpt\\MCP\\mcp_server_toolkit\\build\\index.js"
            ]
        }
    }
}

# Navigate to your MCP directory
export MCP_PATH="/home/codebuilder/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/MCP"
mkdir -p "$MCP_PATH"
cd "$MCP_PATH"

# Clone the repository
git clone https://github.com/franturino/mcp_server_toolkit.git
cd mcp_server_toolkit

npm install

npm run build

# Copy workflow files to your dx-project
cp workflows/* /home/codebuilder/dx-project/.a4drules/workflows/

/home/codebuilder/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/settings/a4d_mcp_settings.json

{
    "mcpServers": {
        "mcp_server_toolkit": {
            "disabled": false,
            "timeout": 600,
            "type": "stdio",
            "command": "node",
            "args": [
                "/home/codebuilder/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/MCP/mcp_server_toolkit/build/index.js"
            ]
        }
    }
}

chmod -R a-w build src
find build src -type d -exec chmod a+rx {} +
find build src -type f -exec chmod a+r {} +

# Make writeable (to edit files)
Get-ChildItem -Path "C:\path\to\build" -Recurse | Set-ItemProperty -Name IsReadOnly -Value $false

# Make read-only (for protection)
Get-ChildItem -Path "C:\path\to\build" -Recurse | Set-ItemProperty -Name IsReadOnly -Value $true

# Make writeable (to edit files)
chmod -R a+w /path/to/build

# Make read-only (for protection)
chmod -R a-w /path/to/build
find /path/to/build -type d -exec chmod a+rx {} +
find /path/to/build -type f -exec chmod a+r {} +

{
    "mcpServers": {
        "mcp_server_toolkit": {
            "disabled": false,
            "timeout": 600,
            "type": "stdio",
            "command": "node",
            "args": [
                "/path/to/mcp_server_toolkit/build/index.js"
            ]
        }
    }
}

cannot be loaded because running scripts is disabled on this system

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

Permission denied: ./setup.sh
./setup.sh: command not found

chmod +x setup.sh update.sh

git is not recognized | git: command not found

npm ERR! code E... | npm ERR! 404 | npm ERR! network ...

error TS... | Error: Cannot find module 'typescript'

File not found: a4d_mcp_settings.json | Invalid JSON | parse error

# Verify repository exists
Test-Path "C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt\MCP\mcp_server_toolkit"

# Verify build files exist
Test-Path "C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt\MCP\mcp_server_toolkit\build\index.js"

# Check configuration
Get-Content "C:\Users\codebuilder\.vscode-server\data\User\globalStorage\salesforce.salesforcedx-einstein-gpt\settings\a4d_mcp_settings.json" | ConvertFrom-Json

# Verify repository exists
test -d ~/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/MCP/mcp_server_toolkit && echo "Directory exists" || echo "Directory missing"

# Verify build files exist
test -f ~/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/MCP/mcp_server_toolkit/build/index.js && echo "File exists" || echo "File missing"

# Check configuration
cat ~/.local/share/code-server/User/globalStorage/salesforce.salesforcedx-einstein-gpt/settings/a4d_mcp_settings.json | jq

## 🛠️ Workflow: Update Defect Status with Validation

This workflow allows you to update the status of a defect, verifying that the provided status is among the allowed values.

### 📥 Input

The workflow accepts two inputs:

- `defect_code`: the defect code to update (e.g., D-123)
- `defect_status`: the status to assign to the defect

Valid statuses include:

New, Re-opened, Sospeso, Need More Information, Non Deployabile, Fix in Process,
Ready to Test, In Deployment, Deploy PREPROD, Ready to test - NEW INTEGRA,
Deploy SUPPORT, Ready to test - PREPROD, Ready to test - SUPPORT,
Closed - End of life, Closed - Duplicate, Closed - Rejected, Closed - Resolved,
Not a Defect, Deploy PROD

### ⚙️ Behavior

- If the inputs are **provided directly in the prompt**, the workflow uses them without asking.
  - Example:
    ```
    /aggiorna-stato-defect.md "D-123" "Ready to Test"
    ```
- If one or both inputs are **missing**, the workflow will prompt the user interactively.
- The workflow **validates the status**:
  - If the status is invalid, it displays an error message with the list of allowed statuses.
  - If the status is valid, it invokes the MCP server `mcp_server_toolkit` using the method `updateDefectStatus`.

### 📤 Output

The workflow creates a task that updates the status of the specified defect, useful for operational management and change tracking.

## 📋 Workflow: Retrieve Multiple Defect Details

This workflow allows you to retrieve details for one or more defects specified by the user.

### 📥 Input

The workflow accepts one input:

- `defect_codes`: a list of defect codes separated by commas (e.g., `D-123,D-456,D-789`)

### ⚙️ Behavior

- If the `defect_codes` input is **provided directly in the prompt**, the workflow uses it without asking.
  - Example:
    ```
    /richiedi-defect-details.md "D-123,D-456"
    ```
- If the input is **missing**, the workflow will prompt the user interactively.
- The workflow loops through each provided code and creates a task to retrieve defect details using the MCP server `mcp_server_toolkit`.

### 📤 Output

For each code provided, the workflow creates a task that invokes the `getDefectDetails` method on the MCP server, useful for analysis, debugging, or tracking.

## 🧾 Workflow: Insert Manual Configuration

This workflow allows you to insert a manual configuration in Salesforce, associated with a specific defect, using the MCP server.

### 📥 Input

The workflow accepts two inputs:

- `defect_code`: the defect code (e.g., D-123)
- `config_note`: the note to associate with the manual configuration

### ⚙️ Behavior

- If the inputs are **provided directly in the prompt**, the workflow uses them without asking.
  - Example:
    ```
    /inserisci-configurazione-manuale.md "D-123" "Configuration required for test environment"
    ```
- If one or both inputs are **missing**, the workflow will prompt the user interactively.
- Once the data is collected, the workflow invokes the MCP server `mcp_server_toolkit` using the method `insertManualConfiguration`.

### 📤 Output

The workflow creates a task that records the manual configuration associated with the specified defect, useful for change tracking and technical management.

## 🧾 Workflow: Create Metadata for Import to Salesforce

This workflow allows you to create a metadata record for importing data into a specific Salesforce table using external codes. It interacts with the MCP server tool `create_metadata_for_import_salesforce`.

### 📥 Input

The workflow accepts the following inputs:

- `defectId` *(optional)*: The ID of the Defect__c record (e.g., `a1B0500000abcdeAAA`)
- `defectName` *(optional)*: The name/code of the defect (e.g., `D-123`)
- `tableName`: The name of the table where records will be imported
- `externalCodesInput`: A comma-separated list of external codes to associate with the metadata

### ⚙️ Behavior

- If the inputs are **provided directly in the prompt**, the workflow uses them without asking.
  - Example:
    ```
    /crea-metadata-import-salesforce.md "D-123" "ImportTable__c" "codeA,codeB,codeC"
    ```
- If one or more inputs are **missing**, the workflow will prompt the user interactively.
- The workflow converts the external codes into an array and invokes the MCP server tool `create_metadata_for_import_salesforce`.

### 📤 Output

The workflow creates a task that inserts a `NDC_XMLObject__c` record in Salesforce, associating it with the specified defect and external codes. The result includes a success or error message returned by the MCP server.

This workflow is useful for automating metadata creation for bulk imports and ensuring traceability through defect linkage.

## 📊 Workflow: Query Metadata by Defect in Salesforce

This workflow allows you to retrieve metadata records associated with a specific defect in Salesforce, using either the defect ID or its code. It interacts with the MCP server tool `query_metadata_by_defect_salesforce`.

### 📥 Input

The workflow accepts the following inputs:

- `defectId` *(optional)*: The ID of the Defect__c record (e.g., `a1B0500000abcdeAAA`)
- `defectCode` *(optional)*: The code of the defect (e.g., `D-123`)

### ⚙️ Behavior

- If the inputs are **provided directly in the prompt**, the workflow uses them without asking.
  - Example:
    ```
    /query-metadata-by-defect.md "D-123"
    ```
- If neither input is provided, the workflow will prompt the user to enter either the defect ID or code.
- The workflow invokes the MCP server tool `query_metadata_by_defect_salesforce`, which:
  - Resolves the defect ID if only the code is provided
  - Queries all metadata records (`NDC_XMLObject__c`) linked to the defect
  - Returns the result in JSON format

### 📤 Output

### 📤 Output

The workflow creates a task that returns the metadata records associated with the specified defect. This is useful for tracking, auditing, or validating import configurations.

---

## Available Workflows

The toolkit includes pre-configured workflows for managing defects and metadata in Salesforce. Each workflow is defined as a Markdown file in the `workflows/` directory.

### 📋 Workflow List

| Workflow File | Purpose | Inputs |
|---------------|---------|--------|
| `retrieve-metadata-of-defect.md` | Query metadata associated with a defect | Defect ID or Code |
| `update-status-defect.md` | Update defect status with validation | Defect Code, New Status |
| `get-detect_details.md` | Retrieve details for one or more defects | Comma-separated defect codes |
| `get-defect_details_by_status_and_developer.md` | Get defects by developer and status filter | Developer Name, Status(es) |
| `insert-confmanuale.md` | Insert manual configuration for a defect | Defect Code, Configuration Note |
| `create_metadata_defect.md` | Create metadata record for Salesforce import | Defect ID/Code, Table Name, External Codes |
| `insert-ExternalCode.md` | Link external codes to defect metadata | Defect Code, External Code List |

### 🔄 Workflow Features

All workflows include:
- ✅ **Interactive prompts** - Request user input if not provided
- ✅ **Input validation** - Check for allowed values
- ✅ **Error handling** - Display clear error messages
- ✅ **MCP integration** - Use the `mcp_server_toolkit` for backend operations

### 📥 Using Workflows

#### In Salesforce Code Builder

```
/workflow-name.md [arg1] [arg2]
```

**Example** (provide all arguments):
```
/update-status-defect.md "D-123" "Ready to Test"
```

**Example** (let workflow prompt for missing arguments):
```
/update-status-defect.md
# Workflow will ask: "Enter Defect Code: ", then "Enter Status: "
```

#### Common Workflow Parameters

**Valid Defect Statuses**:
```
New, Re-opened, Sospeso, Need More Information, Non Deployabile, 
Fix in Process, Ready to Test, In Deployment, Deploy PREPROD, 
Ready to test - NEW INTEGRA, Deploy SUPPORT, Ready to test - PREPROD, 
Ready to test - SUPPORT, Closed - End of life, Closed - Duplicate, 
Closed - Rejected, Closed - Resolved, Not a Defect, Deploy PROD
```

---

## 📞 Support & Documentation

### Getting Help

1. **Check the log files** - Located in the setup script directory
   - `install_YYYYMMDD_HHMMSS.log` - Installation log
   - `update_YYYYMMDD_HHMMSS.log` - Update log

2. **Review the Troubleshooting section above** - Solutions for common issues

3. **Verify your installation** - Use the checklist in Troubleshooting → Verification Checklist

4. **Check file permissions** - Especially important on Unix/Linux systems

### Repository Links

- 📦 **MCP Server Toolkit**: https://github.com/franturino/mcp_server_toolkit
- 🛠️ **Setup Scripts**: https://github.com/franturino/mcp-server-setup
- 🐛 **Report Issues**: Create an issue on GitHub

### Key Features Summary

✅ **Automated Installation** - Single command setup  
✅ **Simple Updates** - Pull latest changes anytime  
✅ **File Protection** - Automatic read-only management  
✅ **Workflow Tools** - Pre-built automation workflows  
✅ **Detailed Logging** - Complete audit trail  
✅ **Salesforce Integration** - Full MCP server support  
✅ **Cross-Platform** - Works on Windows, Linux, and macOS  

---

## 📝 After Installation

### First Steps

1. ✅ Verify MCP server is active in VS Code
2. ✅ Restart VS Code or Code Builder
3. ✅ Check workflows are available (try `/` to see list)
4. ✅ Run a simple workflow to test

### Regular Maintenance

- **Monthly**: Run `update.ps1` or `update.sh` to get latest features
- **As needed**: Check log files for warnings or errors
- **Periodically**: Verify file permissions are correct

### Staying Updated

The update scripts keep you current with:
- Latest bug fixes
- New features and workflows
- Security updates
- Dependency updates

Simply run the update script whenever you want the latest version!

---

**Last Updated**: November 12, 2025  
**Version**: 2.1 (Complete Installation & Usage Guide)