Documentation · MCP · Protocol 2024-11-05 · v2.2.0

Model Context Protocol server

Bridge private Career Context and VitaeGraph to AI coding tools across any project workspace over stateless stdio JSON-RPC 2.0.

Cross-workspace bridge

Stateless Model Context Protocol server.

Connect your private Career Context and VitaeGraph directly to AI coding assistants across any project workspace without copying career files or pasting private notes into chats.

Standalone stdio execution

npx -y vitaecontext-mcp

Or via the package CLI: npx vitaecontext mcp

Why MCP for career facts

Zero Workspace Pollution: No personal Markdown or career files need to be committed to software repositories.
Instant Cross-Project Grounding: Any tool supporting Model Context Protocol can query your career records on demand.
Always Current: Edits to ~/.vitaecontext/career-context.md or ~/.vitaecontext/vitaegraph/ immediately reflect across all chats.
Evidence-Bounded: Exposes structured tools that extract only verified facts for the requested surface.

Architecture & Protocol

Architecture

Stateless stdio server (zero external dependencies)

Transport

stdio JSON-RPC 2.0 with newline framing

Specification

Model Context Protocol (Open Standard)

Package Engine

vitaecontext v2.2.0 (pure ESM)

Discovery Paths

~/.vitaecontext/*-career-context.md, ~/.vitaecontext/vitaegraph/

Core Motivation

Why career context needs a stateless protocol bridge

How an MCP server solves the friction of maintaining accurate, private career facts across developer workflows.

The Multi-Repository Dilemma

Developers routinely work across dozens of git repositories, including open-source packages, client apps, internal microservices, and personal tools. When building or publishing projects, AI coding assistants often need your career context to draft repository READMEs, write author bios, create portfolio entries, or formulate technical accomplishments.

Without an MCP bridge, developers face two bad choices: copy a private career-context.md into every project tree (risking accidental public commits of private data) or manually paste unverified career notes into every chat session.

Private, Single Source of Truth

VitaeContext MCP eliminates this friction by running locally over standard input/output (stdio). It directly connects AI agents to your central ~/.vitaecontext/ home directory on demand.

When you update a role description or record a new achievement in VitaeGraph, every AI assistant in every repository immediately accesses the verified, up-to-date facts on its next query, with zero file copies and zero configuration sprawl.

Client setup

Configure your AI assistant

Claude Desktop

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or ~/.config/Claude/claude_desktop_config.json (Linux)

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

Claude Code

Terminal command

claude mcp add vitaecontext -- npx -y vitaecontext-mcp

Cursor IDE

Settings -> Features -> MCP Servers -> Add new MCP server

Name: vitaecontext
Type: command
Command: npx -y vitaecontext-mcp

Windsurf (Cascade)

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

Roo Code / Cline

cline_mcp_settings.json or roo_code_mcp_settings.json

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

Antigravity CLI

~/.gemini/antigravity-cli/mcp_config.json

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

IBM Bob / watsonx

.ibm/mcp_config.json

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

xAI Grok

.grok/mcp_config.json

{
  "mcpServers": {
    "vitaecontext": {
      "command": "npx",
      "args": ["-y", "vitaecontext-mcp"]
    }
  }
}

Testing

How to test the MCP server

Verify JSON-RPC communication locally using the inspector or direct stdio piping.

Official MCP Inspector

Test tools, resources, and prompts interactively with a web dashboard:

npx @modelcontextprotocol/inspector npx -y vitaecontext-mcp

Open the displayed URL, connect via STDIO, and test tools such as get_career_context.

Command-line stdio piping

Send standard JSON-RPC 2.0 payloads directly via stdin:

printf '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}\n' | npx -y vitaecontext-mcp