Week 4

Skills

Different Models

Connectors (MCP)

Subagents

Practice

Assignment

Back to Track

Connectors (MCP): hands beyond the repo

You've already used one connector: Figma, in Week 1. This page names the mechanism and extends it.

Watch: MCP in Claude Code

https://www.youtube.com/watch?v=kkBFmwkDzdo&list=PLmWCw1CzcFilebjK89WLb5cAvM8K0cLB3&index=8

What MCP is

Model Context Protocol — an open standard for giving the agent access to external services: design files, databases, issue trackers, docs, browsers. A connector is the agent's hands beyond the repository: instead of you copy-pasting data in, the agent reads (and sometimes writes) the service directly.

You can think about MCP like an API but for AI agents.

Connecting

Same two paths as everything in this track:

 — the CLI's view of the same list you saw in Desktop's Connectors settings. Note the "27 tools" next to figma: a connector isn't an abstraction, it's a concrete set of capabilities the agent gains.

/mcp — the CLI's view of the same list you saw in Desktop's Connectors settings. Note the "27 tools" next to figma: a connector isn't an abstraction, it's a concrete set of capabilities the agent gains.

What to connect — and what not to

Connect what the project actually needs: Figma for design work, your database for data tasks, Notion/docs if specs live there. Every connector is also a grant of access: the agent can now act on that account. Three rules from Week 3 carry over:


The HackYourFuture curriculum is licensed under CC BY-NC-SA 4.0 *https://hackyourfuture.net/*

CC BY-NC-SA 4.0 Icons

Built with ❤️ by the HackYourFuture community · Thank you, contributors

Found a mistake or have a suggestion? Let us know in the feedback form.