This starter template is designed to leverage Context Engineering—a method of systematically curating information to maximize the effectiveness of AI coding agents like Cline.
"Context engineering is curating what the model sees so that you get a better result."
— Bharani Subramaniam (via Martin Fowler)
(Satirical / Emerging Term)
A tongue-in-cheek title for the modern developer whose primary "coding" activity has shifted from writing complex syntax to writing structured Markdown files and natural language rules.
You might be a Markdown Engineer if:
- 📂 Your "codebase" is mostly
.mdfiles instructing an AI. - 🧹 You spend more time curating context than debugging logic.
- 🗣️ You use English as your primary programming language.
Definition synthesized with the help of Gemini. 🤖
The short answer: I was asked to dig up an old project, saw the code, and realized life is too short to refactor spaghetti manually.
The long answer:
This repository wasn't born out of a desire to replace developers. It was born out of pragmatic laziness and a need for quality control.
I had to revisit a legacy codebase where the code was... let's just say "adventurous." I frankly didn't have the energy to type out fixes line-by-line. I wanted a tool to do the heavy lifting so I could focus on the architecture (and my sanity).
We often hear that AI writes bad code. But let's be honest: humans write bad code, too.
I’ve met plenty of developers who wear the badge of being "Otodidak" (Indonesian for autodidact or self-taught). While self-learning is admirable, I've seen it used as a shield to justify skipping best practices or ignoring foundational engineering principles.
I’ve cleaned up the holes they left behind:
- The developer who proudly deployed at 1 AM, broke production, and had no idea why.
- The time an image kept getting mysteriously "cut off" in the UI, leaving the original dev baffled until a reliable engineer had to step in and purge the mess.
At that point, does it matter if the bad code came from a hallucinating LLM or a reckless human?
I accept that I am just a normal engineer. I’m not claiming to be a 10x developer, and neither is this AI.
The goal of this template is Augmentation, not Abdication:
- Generate: Let the AI write the boilerplate, the refactors, and the tedious parts.
- Supervise: A human must review the output. You are the adult in the room. You are the one who has to answer to the client when it breaks.
Use this template to move fast, but don't take your hands off the wheel. 🚗💨