OPEN TO FULL-TIME ROLES (Available to start immediately. Remote preferred worldwide.)// STACK:WORDPRESS & WOOCOMMERCEREACT & NEXT.JSTYPESCRIPTWEB PERFORMANCEAI INTEGRATIONSAUTOMATION (N8N, MAKE)// PROTOCOL: MCP · JSON-RPC 2.0// STATUS: PRODUCTION READY
OPEN TO FULL-TIME ROLES (Available to start immediately. Remote preferred worldwide.)// STACK:WORDPRESS & WOOCOMMERCEREACT & NEXT.JSTYPESCRIPTWEB PERFORMANCEAI INTEGRATIONSAUTOMATION (N8N, MAKE)// PROTOCOL: MCP · JSON-RPC 2.0// STATUS: PRODUCTION READY
Back to all projects
// mcp★ Featured System

Agent-Ready Portfolio

An agent-ready portfolio with MCP, llms.txt, and an interactive terminal for AI assistants.

[01] Problem Statement & Context

The Challenge

Recruiters and hiring tools increasingly run AI agents to screen candidates before a human ever opens a resume, but almost every developer portfolio is built for human eyes only — flat HTML an agent has to scrape and guess at, with no reliable way to ask it a direct question and get a structured answer back.

[02] Technical Architecture & Solution

Engineering Delivery

Built the portfolio itself as the proof: two parallel layers reading from the same source-of-truth data. The human layer is an Astro 5 + React 19 site with a hero terminal that runs real commands (including a hidden multi-chapter narrative easter egg) with a typewriter reveal. The agent layer is a stateless MCP server speaking JSON-RPC 2.0 over HTTP at /api/mcp, exposing read-only tools (resume, projects, capabilities, availability, and the hidden manifest) plus one tool with a side effect (email contact via Resend). Any MCP-compatible client — Claude, Cursor, ChatGPT — connects directly and queries the exact same data the human site renders, with zero scraping.

[03] Technology Stack & Utilities
Astro 5React 19TypeScriptTailwind CSS v4MCP · JSON-RPC 2.0Vercel Edge FunctionsResend