docs+menu: element icons in the field Magic list, and kits.md's Chip column states unbuilt intent as fact

[mtklein]

Two related gaps left after #53 put element icons on the field Blitz page.

1. The field Magic/Lore list and the esper detail page show no element. Spells carry elements and the battle magic list already shows them, so the field is the inconsistent surface. Not blocked on space — the row is a 7-cell name plus a blank plus 2 digits at columns 3 and 16, leaving 13-15 and 26-29 free — but every MagicName already opens with a school glyph, so this is a two-glyphs-per-row design question, not a plumbing one. It also goes through vanilla's shared drawer _c34fc4, which has several callers, and the field magic page has no render test to fail first against.

2. kits.md's Chip column states unbuilt intent as fact (found while auditing #54): AutoCrossbow "×4" is really one hit per body — breadth, not rate, so exactly one chip against a boss — and Bum Rush "×8", Pummel "×2" and Drill "×2" are all ×1 today. Correct the column to describe what ships, and mark intent as intent. DESIGN.md:147's "Cyan's Flurry" also wants the FF3-US name, Quadra Slam.

Related: ot6_class.asm:185-192 carries retranslated names as code-comment parentheticals (; dispatch (fang), ; quadra slam (flurry — chips per hit)), which is the thing #51 was clearing.

[mtklein]

Partly done in the tech-debt pass (81521ce). kits.md's Chip column is corrected — dated correction at the head of "The constrained three" with the real three-ability table, plus pointer notes on Sabin's table and Edgar's "first shield shredder" bullet. AutoCrossbow is called out as a different kind of error than the others: it is one hit per body, so breadth rather than rate, and exactly one chip against a boss.

DESIGN.md:147's "Flurry" was already fixed by #51's sweep (4126360) — my note here was wrong about that — though the same sentence was false for a second reason (it named AutoCrossbow as multi-hit) and got a correction anyway.

Still open: element icons in the field Magic list and on the esper detail page. Unchanged from the description above — it is a two-glyphs-per-row design question rather than a space problem, since every MagicName already opens with a school glyph, and it goes through vanilla's shared drawer _c34fc4 with several callers. The field magic page still has no render test to fail first against.

Also still open, moved here from the docs sweep: ot6_class.asm:185-192's retranslated parentheticals — see #73 §6.