I made two skill files for Claude that turned my engineering ideas into buildable steps
Disclaimer: What you are seeing is a polished version of my draft, polished by Claude.
I made two skill files for Claude
- is Project Intelligence Layer (IL). You tell it your goal, it breaks it into steps in real build order — not textbook order, actual order.
Four steps: understand, explain, break into parts, produce output.
The output is ready to use directly MATLAB simulation prompt, Keil C structure, Proteus steps, whatever your target platform is.
- Aizen Tutor Depth calibrator with 9 levels inspired by chess ranking. It controls how deep the explanation goes at each step. So the loop doesn't over-explain what you already know.
Together they form a ReAct loop — Reason → Act → Observe → Correct → Repeat. But with depth control at each step.
Tested on: MATLAB simulations, college report making, image prompting in Gemini Nano, AI music, Google AI Studio coding prompts (reduced hallucination), cv making, self test
i think both skill is works on any cause only need to calibrate accrding use caus
anyone tell me how to solve thos problems
problems:
currently i using both as chatbot i try to switch to agentic
Both skills consume a lot of tokens in one chat
i not using this in any paid models