akshaymemane/asd-ste100-teach-skill

A Claude Code skill that turns one technical topic — a book chapter, a paper, a system-design concept — into four learning artifacts that share a single core example:
| Artifact | What it is |
|---|---|
| 01-explanation.md | Simplified Technical English explanation: terms defined before use, trade-offs, common mistakes, five test questions |
| 02-diagrams.md | Mermaid architecture, data flow, algorithm, sequence, failure/recovery, and alternatives diagrams |
| 03-interactive/index.html | Standalone simulation: parameter sliders, play/pause/reset/step, failure injection, tooltips, knowledge checks |
| 04-video/ | Manim source plus a rendered MP4. Length follows the content — a subtitles-only cut runs 60–120 s |
This skill extends danyuchn/asd-ste100-skill, which owns the controlled-language rules. This one owns the learning-pack workflow and delegates every line of prose to it.
The audience it writes for: an experienced backend engineer who is new to the specific topic. Flat language, exact mechanics, no accuracy traded for simplicity.
Install
Both skills are needed. User-level install, so they work in every directory on the machine:
Optional slash command:
Optional narration (macOS): NARRATE=1 makes every caption speak itself, then scripts/narrate.py mux attaches the track in sync.
Check the toolchain:
It reports python3, the STE linter, npx, manim, and ffmpeg, and prints the install command for anything missing. Manim needs brew install cairo pango pkg-config && pip3 install manim on macOS.
Use
Or plain language — the skill triggers on "teach me X", "explain chapter N of …", "make me an interactive lesson", "build a visual explainer".
The skill stops and asks for source material rather than inventing chapter content.
Layout
License