yiibulisted

# yiibu — Video Post-Production All shell commands in this document run from the skill root (the directory containing this SKILL.md), wherever it is installed. Automated pipeline for turning raw selfie recordings into polished 9:16 short-form videos. ## Division of labour **The user shoots. You select.** They hand over a folder; deciding which shots earn a place, in what order, and how long the result runs *is* the edit. Do not ask them to pick clips — that is the work. Do ask when a fact on screen is unreadable, because they were there and you were not. New to this skill or a new machine: `python3 doctor.py` (needs only ffmpeg, Pillow, numpy — the font is bundled, everything else degrades cleanly). ## Why so much of this is code and not advice The same footage should come out at the same STANDARD whoever is driving — caption style, audio targets, cover geometry, honest reporting of what is unfinished. Not the same cuts or the same rhythm; those are judgement and should differ. The standard should not. Prose cannot deliver that. Every rule left as a sentence in this file gets re-derived by whoever reads it next, and the record is that they re-derive it differently — or skip it and report success. So the rules that matter live in three places that do not depend on anyone agreeing with them: | mechanism | what belongs there | why | |---|---|---| | `house_style.json` + `gates.py` | measurable, stable threshold, defect survives human review | a number outside the thresho