The 114 Build Plan

Why these nine, and not nine equal ones

Balancing by hours alone would cut through the middle of al-Ḥawāmīm and separate al-Anfāl from al-Tawba. The tradition already divides the muṣḥaf — al-Sabʿ al-Ṭiwāl, al-Miʾūn, al-Mathānī, and the three classes of al-Mufaṣṣal — and it divides it where length, register and subject actually change. Cutting there costs a little balance and buys consistency, which is the thing at risk.

Three rules drove the boundaries: a recognised family is never split (40–46 all open Ḥā Mīm); sūrahs read as a pair stay together (al-Anfāl and al-Tawba, which no basmala divides); and a window holds one register where it can, so Madanī law does not sit beside Makkan eschatology in the same session.

Three at a time — the order matters

Windows in the same wave cannot see each other, so put the vocabulary-setting ranges first. Wave 1 covers the remaining long sūrahs, the Ḥā Mīm family and the eschatological register — between them they settle most of the terms the later waves will need.

After the last wave, run one reconciliation pass: a single window reads the glossary and greps every settled term across all 114 readings. That is the only step that can catch what the per-sūrah gates cannot.

The prompt for each window

Set the model first — a prompt cannot do it. In each window run /model claude-opus-5 before pasting. There is nothing to gain from a cheaper tier here: the mechanical steps are three commands and cost almost no tokens, so a downgrade buys speed only on the readings, which is the one part no rebuild can regenerate.

Then paste this, with the numbers for that window filled in from the table above:

You are WINDOW 1 of 9 building the Qur'an surah pages.
Your range is surahs 3-7 - 5 surahs, 867 ayat. No other window
touches them.

FIRST, before any reading: read 05-READINGS/GLOSSARY.md and hold to it.
It is the controlled vocabulary for all nine windows. When you settle a
term it does not carry, append a row. Never edit a settled row - put an
objection under ## Disputed and leave it for Waleed.

Then work ONE SURAH AT A TIME, in order, all five steps of the skill, and
publish it before starting the next. A finished surah should be finished -
do not batch the publishing to the end.

WHERE TO SPEND THOUGHT. The agreed readings and the hadith picks are the
entire job; everything else is three commands. Run --data, --build and
--audit without deliberation. Think hard about: what the commentaries
SHARE for this ayah, whether a term is already settled in the glossary,
and whether the hadith picks carry a cross-tradition convergence. If a
gate fails, fix the input - never the gate.

WHEN THE RANGE IS DONE, report: surahs built, total ayat, every gate that
fired and why, and every term you added to the glossary.

Start now with /surah-purge 3

The skill carries the rest — the five steps, the consensus rules, the batch tooling for long sūrahs, and every gate. Do not restate it in the prompt; a longer prompt only competes with the skill.

All 114, in build order