explainer-layer-stacklisted
Install: claude install-skill Safiullah-Rahu/claude-video-skills
# Layer-stack — presenter + piece-by-piece system diagram reel
> `ENGINE` = `~/.claude/skills/explainer-engine/engine` (the **explainer-engine** skill: renderer, voice, score, QA). Rendered sample of this skill: `the sample film (the render is not included in this repository)` (source in `assets/sample/`).
A product that people think is "one thing" is rebuilt layer by layer, each layer bolted on after something
concrete broke. One warm canvas, one static narrator character at the left, and a flat ink-outline prop world on
the right that hard-cuts to a new arrangement every ~6 s while a small **layer stack** grows row by row. Builds
land on every stressed word (~1.5 per second), failures end in a red rubber stamp, and the reel ends on a callback:
the very first failure returns through one small human edit. Built from frame analysis of 1 source reel
(references/sources.md, references/analysis.md). **Original work only**: design your own presenter (never a known
character), your own mascot, brand and script; no channel marks.
**Format: vertical 1080×1920, 30 fps, 115–140 s** (`<html data-size="1080x1920">`). For 16:9: `data-size="1920x1080"`,
presenter at x 40–300 (feet at 92 % H, height 60 % H), diagram in x 360–1880, captions centred at 86 % H, chapter tag
at the same 12 % / 7.6 % offsets; scale everything else by H.
## 1. Signature traits
- **Warm canvas `#faf8f4` that never changes**, no texture, no grain, no gradients; flat fills, 3 px ink outlines,
hard grey offset s