A collection of Model Context Protocol servers for databases, remote access, automation, and developer tools. Each server works standalone with any MCP-compatible client.

All nine servers run on MCP SDK 2.x and speak protocol revision 2026-07-28, while still serving 2025-era clients from the same process. What that buys you, across the board:

• Structured output — every tool declares an outputSchema and returns structuredContent, so results are typed JSON instead of formatted text a client has to parse back
• Real tool errors — failures come back with isError, not as text that looks like an answer
• Confirmation before damage — destructive operations (write queries, rm -rf, killing a process, overwriting a file, wiping a chat room) ask the user first, and skip the question when there is nothing to lose
• Progress and cancellation — long operations report progress; cancelling a call cancels the work behind it
• Resources and prompts — schemas, logs and histories are readable as resources, with ready-made prompts for the common investigations
• Tool annotations — readOnlyHint / destructiveHint / idempotentHint on every tool

See MIGRATION-V2.md for the per-server record of what changed.

./scripts/run-tests.sh

Each server has its own suite: pytest for the Python servers, npm test for the Node ones. The PostgreSQL and Oracle suites need a live database and skip themselves when one is not configured — docker-compose.yml brings up both, and the script's header has the environment variables to export.

See RELEASING.md. scripts/build-release.sh produces every distributable and scripts/verify-release.py installs each one into a throwaway environment and starts it — the check that catches packaging bugs the test suites cannot see.

Install any server with a single command — no virtual environment needed:

# Python servers
uvx mapeg-postgres-mcp
uvx mapeg-oracle-mcp
uvx n8n-chatbot-mcp
uvx docusaurus-mcp
uvx gemini-reviews-mcp
uvx mcp-server-ssh
uvx agent-chat-mcp

# Node.js servers
npx -y asger-terminal-mcp
npx -y gib-api-mcp

claude mcp add postgres \
  -e DB_HOST=localhost \
  -e DB_PORT=5432 \
  -e DB_NAME=mydb \
  -e DB_USER=postgres \
  -e DB_PASSWORD=secret \
  -- uvx mapeg-postgres-mcp

Add to your config file:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "postgres": {
      "command": "uvx",
      "args": ["mapeg-postgres-mcp"],
      "env": {
        "DB_HOST": "localhost",
        "DB_PORT": "5432",
        "DB_NAME": "mydb",
        "DB_USER": "postgres",
        "DB_PASSWORD": "secret"
      }
    },
    "oracle": {
      "command": "uvx",
      "args": ["mapeg-oracle-mcp"],
      "env": {
        "ORACLE_CONNECTION_STRING": "User Id=myuser;Password=mypass;Data Source=host:1521/service"
      }
    },
    "my-chatbot": {
      "command": "uvx",
      "args": ["n8n-chatbot-mcp"],
      "env": {
        "N8N_CHATBOT_URL": "https://n8n.example.com/webhook/my-bot/chat"
      }
    }
  }
}

Add to ~/.cursor/mcp.json using the same format as Claude Desktop.

Add to Windsurf MCP config using the same format as Claude Desktop.

Add to your VS Code settings (JSON):

"mcp": {
  "servers": {
    "postgres": {
      "type": "stdio",
      "command": "uvx",
      "args": ["mapeg-postgres-mcp"],
      "env": {
        "DB_HOST": "localhost",
        "DB_PORT": "5432",
        "DB_NAME": "mydb",
        "DB_USER": "postgres",
        "DB_PASSWORD": "secret"
      }
    }
  }
}

Add to ~/.gemini/settings.json:

{
  "mcpServers": {
    "postgres": {
      "command": "uvx",
      "args": ["mapeg-postgres-mcp"],
      "env": {
        "DB_HOST": "localhost",
        "DB_PORT": "5432",
        "DB_NAME": "mydb",
        "DB_USER": "postgres",
        "DB_PASSWORD": "secret"
      }
    }
  }
}

Add to ~/.copilot/mcp-config.json:

{
  "mcpServers": {
    "postgres": {
      "command": "uvx",
      "args": ["mapeg-postgres-mcp"],
      "env": {
        "DB_HOST": "localhost",
        "DB_PORT": "5432",
        "DB_NAME": "mydb",
        "DB_USER": "postgres",
        "DB_PASSWORD": "secret"
      }
    }
  }
}

Add to ~/.codex/config.toml:

[mcp_servers.postgres]
command = "uvx"
args = ["mapeg-postgres-mcp"]

[mcp_servers.postgres.env]
DB_HOST = "localhost"
DB_PORT = "5432"
DB_NAME = "mydb"
DB_USER = "postgres"
DB_PASSWORD = "secret"

# Python servers
cd mapeg-postgres-mcp
pip install -e .

# Node.js servers
cd asger-terminal-mcp
npm install

Database MCP servers include a built-in query_logger module that writes to a shared SQLite database at ~/.local/share/mapeg-mcp/query_logs.db. Each log entry includes:

• db_identifier — which database (e.g. localhost:5432/mydb)
• workspace_path — which project directory

The get_query_history tool lets agents review their own recent queries, scoped to their database and workspace.

Monitor all database queries in real-time:

cd query_dashboard
pip install flask
python app.py
# Open http://localhost:5555

Supports filtering by server type, database, and workspace.

• Store credentials in environment variables, not in code
• Use READ_ONLY=true for database servers when write access isn't needed
• Prefer SSH key-based authentication over passwords
• All .env files are gitignored

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-feature)
3. Commit your changes
4. Push and open a Pull Request

• MCP Documentation
• MCP Specification
• Python MCP SDK
• TypeScript MCP SDK

MIT