DEV Community: 一郎支

I Built a 'McDonald's Persona Card' Skill With Their MCP Server (And You Can Too)

一郎支 — Sun, 11 Oct 2026 00:28:35 +0000

TL;DR

McDonald's China shipped an official MCP server. I used it (through WorkBuddy) to build a zero-setup "persona card" Skill that turns your taste, budget, and mood into a fun 麦门人格卡 (McDonald's persona card) with meal recommendations. It's open source, runs with one command, and it's entered in the McDonald's Developer Innovation Challenge.

If you find it fun or useful, a ⭐ on the repo genuinely helps it rank: https://github.com/Zhiyilang074811/mcd-persona-card

What it does

Input your flavor preferences, budget, and scene, and it generates a personalized "麦门人格卡".
It pairs the persona with a recommended combo and a money-saving bundle.
Ships with a one-command demo and clear MCP integration docs, so even non-developers can reproduce it.

How to run

python mcd_persona.py --demo

No API key required for the demo. The full version wires into the official McDonald's MCP server (instructions in MCP_INTEGRATION.md).

How it uses the McDonald's MCP

The project calls the official MCP tools to fetch real menu, pricing, and promotion data, then layers a lightweight persona engine on top. Details and the call flow live in MCP_INTEGRATION.md.

Why I built it

Most McDonald's helper bots are just calculators. I wanted something that feels personal — a card you'd actually share. Built entirely with WorkBuddy, the official partner of the challenge.

Try it & star it

Repo: https://github.com/Zhiyilang074811/mcd-persona-card

Stars help a small open-source project get seen in the challenge leaderboard. If you liked the idea, hit ⭐ — and feel free to fork or open an issue with your own persona ideas.

GardenPlanner AI: an offline-first gardening planner for Hacktoberfest 2026

一郎支 — Sun, 11 Oct 2026 00:26:41 +0000

GardenPlanner AI is an offline-first gardening planner built for the Hacktoberfest 2026 Open-Source AI Challenge (theme: Touch Grass). The idea is simple: use open-weight models to help people step away from screens and into nature, not pull them deeper in.

Most gardening apps send your location and garden data to a cloud and lock the brain behind a paid API. This tool does the opposite. It runs 100% offline on pure Python standard library, no network, no account, no cost. It optionally rewrites the plan with a local Ollama model; without it, the built-in rule engine still produces a complete plan. Your city and garden data never leave your machine, and the output is a hands-on planting checklist you act on outside.

How to run it:

python main.py --city "Shanghai" --month 10 --space balcony --level easy

It returns a personalized monthly planting plan with companion plants and tips, and flips seasons for the Southern Hemisphere automatically. You can also enhance it with a local model:

ollama pull llama3.2
python main.py --city "London" --month 3 --ollama llama3.2

The repository is here: https://github.com/Zhiyilang074811/hacktoberfest-2026-ai

Contributions are welcome for Hacktoberfest: more city climate data, an expanded plant database, pest reminders, or web/mobile wrappers. Look for the hacktoberfest-accepted label. Let's touch grass, with open source.

probe2

一郎支 — Sun, 11 Oct 2026 00:07:12 +0000

probe2 body

probe

一郎支 — Sun, 11 Oct 2026 00:05:29 +0000

probe

GardenPlanner AI offline gardening planner

一郎支 — Sat, 10 Oct 2026 23:58:19 +0000

Intro about Touch Grass. See https://github.com/Zhiyilang074811/hacktoberfest-2026-ai for code.

python main.py --city Shanghai

Test from API

一郎支 — Sat, 10 Oct 2026 23:56:58 +0000

Hello world. This is a test.

DEV Community: 一郎 支

I Built a 'McDonald's Persona Card' Skill With Their MCP Server (And You Can Too)

TL;DR

What it does

How to run

How it uses the McDonald's MCP

Why I built it

Try it & star it

GardenPlanner AI: an offline-first gardening planner for Hacktoberfest 2026

probe2

probe

GardenPlanner AI offline gardening planner

Test from API

DEV Community: 一郎支