Skip to main content

What is MCP (Model Context Protocol)?

Model Context Protocol (MCP) is an integration system that extends Verdent’s capabilities by connecting to external tools, data sources, and services. Core Concept: MCP enables Verdent to interact with systems beyond its built-in tool set, allowing seamless integration with:
  • External APIs and web services
  • Database systems
  • Cloud platforms (AWS, Azure, GCP)
  • Project management tools
  • CI/CD pipelines
  • Custom internal tools
Architecture: MCP servers act as bridges between Verdent and external systems. Each MCP server exposes specific capabilities that Verdent can invoke during coding sessions, similar to built-in tools like file_read or bash. Configuration Storage: MCP server configurations are stored in ~/.verdent/mcp.json and managed through the Settings menu.

Benefits of MCP Integration

Extended Capabilities

Access External Systems: Connect Verdent to tools and services not available through built-in functionality. Query databases, interact with cloud services, or integrate project-specific APIs without leaving your coding workflow. Custom Tool Development: Create MCP servers tailored to your project’s unique requirements. Build specialized integrations for internal tools, proprietary APIs, or domain-specific systems.

Interoperability

Ecosystem Leverage: Tap into the growing MCP ecosystem. Community-developed MCP servers provide pre-built integrations for popular tools and services. Standardized Protocol: MCP provides a consistent interface for tool integration, ensuring new capabilities integrate smoothly with Verdent’s existing tool system.

Workflow Automation

Seamless Integration: MCP servers become part of Verdent’s tool set. The agent automatically uses MCP tools when appropriate, or you can explicitly invoke them for specific tasks. Context-Aware Operations: MCP integrations understand your codebase context, enabling intelligent interactions with external systems based on current development state.

Setting Up MCP Servers

Configuration Process

Access MCP Settings:
  1. Open Settings menu in Verdent
  2. Navigate to MCP Servers section
  3. View or modify ~/.verdent/mcp.json configuration
Configuration File: MCP servers are configured through the ~/.verdent/mcp.json file, which defines server connections, authentication, and integration parameters. See the MCP Configuration section below for the full schema.

Configuration Management

Adding Servers: Configure new MCP servers by adding entries to mcp.json with connection details and authentication credentials. Multiple Servers: Verdent supports multiple concurrent MCP servers, allowing integration with several external systems simultaneously. Testing Connections: Validate MCP server configurations to ensure successful connections before using them in development workflows.

Server Status

Monitor connection status in the MCP settings panel:
StatusDescription
SuccessServer connected and tools available
FailedConnection error (click to view details)
LoadingServer initializing
Refreshing Tools: Click the refresh button to rediscover available tools from a server after configuration changes or connection issues.

MCP Configuration

Basic Setup

Location: .mcp.json in project root or ~/.verdent/mcp.json (global)
{
  "mcpServers": {
    "postgres": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-postgres",
        "postgresql://localhost:5432/myapp_dev"
      ]
    }
  }
}
Configuration Keys:
  • mcpServers - Required top-level key for MCP configuration
  • command - Executable to run (typically npx for Node.js packages)
  • args - Arguments passed to the command (package name, connection strings, etc.)
  • env - Environment variables for authentication/configuration
Learn More About MCP:

MCP Plugins and Extensions

Available Plugins

MCP plugins provide pre-built integrations for common tools and services. Browse the community registry at mcp.so/servers or the official MCP servers repo for current availability. Common Plugin Categories:
  • Database Integrations: PostgreSQL, MySQL, MongoDB query tools
  • Cloud Services: AWS, Azure, GCP API interactions
  • Developer Tools: GitHub, GitLab, Jira, Linear integrations
  • CI/CD Platforms: Jenkins, GitHub Actions, CircleCI
  • Monitoring Services: Datadog, New Relic, Sentry
  • Communication: Slack, Discord, Microsoft Teams notifications

Plugin Installation

Installation Methods: Plugins are installed and configured through the MCP Settings interface, typically requiring:
  • Plugin identifier or source
  • Configuration parameters
  • Authentication credentials (if applicable)
Dependency Management: MCP plugins may have dependencies on external services or libraries. Installation processes handle dependency resolution automatically when possible.

Custom MCP Server Development

Creating Custom Servers

Advanced users can create custom MCP servers to integrate proprietary tools or domain-specific systems. Development Approach: Custom MCP servers implement the MCP protocol specification, exposing tool capabilities that Verdent can invoke. Refer to the Model Context Protocol specification for implementation details and SDKs. Use Cases for Custom Servers:
  • Internal API Integration: Connect to company-specific APIs and services
  • Legacy System Bridges: Enable Verdent to interact with legacy infrastructure
  • Specialized Data Sources: Access proprietary databases or data warehouses
  • Workflow Automation: Trigger custom build, deployment, or testing pipelines

Integration Patterns

Workspace-Level Configuration: MCP servers can be configured at workspace level for project-specific integrations, enabling teams to share MCP configurations through version control. Tool Interoperability: MCP-provided tools integrate seamlessly with Verdent’s built-in tools (file_read, bash, spawn_subagent), enabling complex workflows combining multiple tool types.

Desktop-Specific MCP Features

Per-Project Configuration

Configuration LevelLocationScope
Global~/.verdent/mcp.jsonAll projects
Project.mcp.json in project rootSingle project
WorktreeInherits from projectInherits project config

Multi-Worktree MCP

MCP servers work across worktrees:
  • Shared Servers: Base workspace MCP config shared by all worktrees
  • Isolated Operations: Each worktree maintains separate MCP session state
  • Parallel Access: Multiple worktrees can use same MCP servers concurrently

Project Switching

When switching projects:
  • Current project’s MCP servers are loaded
  • Global MCP servers remain available
  • Previous project’s MCP connections are maintained in background

MCP Workflow Examples

Database Query Integration

Scenario: Query production database for debugging data issues. MCP Approach:
User: "Check the orders table for customer ID 12345"
Verdent: [Uses MCP database server to execute safe read-only query]
        Results: 3 orders found for customer 12345
        [Displays formatted results]
The MCP database server handles authentication, query validation, and result formatting.

Cloud Service Interaction

Scenario: Check AWS Lambda function logs for errors. MCP Approach:
User: "Show me recent errors from the payment-processor Lambda function"
Verdent: [Invokes MCP AWS integration]
        [Retrieves CloudWatch logs for specified function]
        [Filters for ERROR level entries]
        [Presents formatted error summaries]
MCP AWS server manages authentication, API calls, and log parsing.

CI/CD Pipeline Monitoring

Scenario: Check build status before deploying changes. MCP Approach:
User: "Is the master branch build passing?"
Verdent: [Queries MCP CI/CD integration]
        Build #245: ✓ Passed (2 minutes ago)
        All tests passed, ready for deployment
MCP CI/CD server tracks pipeline status and provides real-time feedback.

Security Considerations

Authentication & Authorization

Credential Management: MCP servers handle authentication to external services. Credentials are stored securely in MCP configuration and never exposed in chat history or code. Permission Scoping: Configure MCP servers with minimal necessary permissions. Use read-only access where possible, and restrict destructive operations.

Safe Operations

Command Review: Operations involving MCP tools appear in the conversation history, providing visibility and control. Audit Logging: MCP server operations appear in Verdent’s chat history, creating an audit trail of external system interactions.

Troubleshooting MCP Integration

Common Issues

Connection Failures:
  • Verify network connectivity to MCP server
  • Check authentication credentials in mcp.json
  • Ensure MCP server is running and accessible
  • Validate configuration file syntax
Permission Errors:
  • Confirm API keys or credentials have required permissions
  • Check service quotas or rate limits
  • Verify firewall rules allow MCP traffic
Tool Invocation Issues:
  • Ensure MCP server exposes expected tools correctly
  • Check tool parameter formats match server expectations
  • Review MCP server logs for error details

Getting Help

Support Channels: Diagnostic Information: When reporting MCP issues, provide:
  • MCP server type and version
  • Configuration details (sanitized of credentials)
  • Error messages from chat history
  • MCP server logs (if accessible)

FAQs

MCP server operations consume credits based on the model processing the results. The MCP server itself doesn’t use credits directly.
No. MCP servers are configured at the user level, not per project. All configured servers are available across all your projects.
Credentials in mcp.json are stored locally on your machine and never sent to Verdent servers. Use environment variables (${VAR_NAME}) for sensitive values.
MCP servers start when needed and run during your session. They don’t consume resources when Verdent is closed.
MCP servers have the permissions you configure. Use read-only access where possible and restrict destructive operations.
Ask Verdent to use a tool from the MCP server. Successful responses indicate the connection is working. Check the Troubleshooting section if issues occur.