diff --git a/TODO.all-sys/INDEX.md b/TODO.all-sys/INDEX.md
index 391367d..23d825f 100644
--- a/TODO.all-sys/INDEX.md
+++ b/TODO.all-sys/INDEX.md
@@ -3,8 +3,12 @@
 Terminal: every identified transliteration system has an ISC map in the
 registry, tested from its own source document, live on the site.
 
-Status line: 318 done / ~45 open. Sources: all local
+Status line: 339 done / ~7 open (owner-gated). Sources: all local
 (interscript-private-references/alalc/ — 92 docs; din/, bsi/ present).
+2026-11-09: the ALA-LC batches 02-05 landed (maps #263-#266, 24 new
+maps, ~1,292 new embedded tests); the corpus is 428 system maps; the
+only open work orders left are owner-gated (08 source acquisition,
+10 B3 ship, 11 urd-oracle license, 12 misc gates).
 
 Rules (binding): visual transcription from scans is sanctioned; every
 map carries tests from its own document's examples; one PR per batch;
@@ -14,23 +18,33 @@ versions are owner decisions.
 - [x] 01-ala-lc-batch-1.md — Brahmic/SE-Asian, 14 systems — LANDED:
       all 14 in the corpus (cjm carries the Cham system; corpus check
       2026-11-09)
-- [ ] 02-ala-lc-batch-2.md — Arabic-script, 13 systems — IN FLIGHT
-      (agent campaign 2026-11-09): ara/aze/fas already in the corpus;
-      kazakh, kurdish, ladino, malay/jawi, sindhi, uighur, urdu,
-      pushto, ottoman sources present; moghol has no source PDF
-      (record, don't encode)
-- [ ] 03-ala-lc-batch-3.md — Cyrillic, 4 systems — IN FLIGHT (agent
-      campaign 2026-11-09) for chu (church slavic) + rue (rusyn),
-      the two with sources; bak/chv/udm/kv/sah have no source PDF
-      (record, don't encode); bel/srp already in the corpus
-- [ ] 04-ala-lc-batch-4.md — East Asian, 2 systems — IN FLIGHT (agent
-      campaign 2026-11-09): chinese (Wade-Giles syllabary) + japanese
-      (McCune-Reischauer kana, M-remainder noted); kor already in the
-      corpus
-- [ ] 05-ala-lc-batch-5.md — misc, 10 systems — IN FLIGHT (agent
-      campaign 2026-11-09) for the 8 with sources (armenian, coptic,
-      deseret, cherokee, manchu, tamazight, tamashek, vai); nko +
-      tod-oirat have no source PDF (record, don't encode)
+- [x] 02-ala-lc-batch-2.md — Arabic-script, 13 systems — LANDED maps
+      #266 (2026-11-09): kazakh, kurdish, ladino (Hebr!), judeo-arabic
+      (jba — source found in the collection beyond the WO's list),
+      malay/zsm, jawi/jav-Arab, sindhi, urdu, uighur, pushto, ottoman
+      (1997 edition per the ALA-LC system; source undated, flagged).
+      681 embedded tests. NOT encodable: moghol (no source PDF).
+      ara/aze/fas were already in the corpus.
+- [x] 03-ala-lc-batch-3.md — Cyrillic, 4 systems — LANDED maps #263
+      (2026-11-09): church slavic (chu-Cyrl-2022, 48 letters × case,
+      yuses/ksi/psi/fita/izhitsa/both uks) + rusyn (rue-Cyrl-2013, tied
+      digraphs, Vojvodinian-i variant documented). 176 embedded tests.
+      NOT encodable: bashkir, chuvash, udmurt, komi, yakut/sakha (no
+      source PDFs). bel/srp were already in the corpus.
+- [x] 04-ala-lc-batch-4.md — East Asian, 2 systems — LANDED maps #265
+      (2026-11-09): japanese (jpn-Hrkt-2022, both MR charts A+B,
+      deterministic rules; kanji readings = documented M remainder) +
+      chinese (zho-Hani-2012, PINYIN-based — the WO's Wade-Giles
+      premise applies to the pre-2000 superseded tables; 321 attested
+      readings) + the document's complete Wade-Giles→Pinyin
+      correspondence table as zho-Latn-Latn-2012 (all 397 rows).
+      214 embedded tests. kor was already in the corpus.
+- [x] 05-ala-lc-batch-5.md — misc, 10 systems — LANDED maps #264
+      (2026-11-09): armenian 2022, coptic 2014, deseret
+      (und-Dsrt-2016, language-neutral script — no corpus precedent,
+      flagged for owner), cherokee 2012, manchu 2012, moroccan
+      tamazight 2012, tamashek 2013, vai 2010. 221 embedded tests.
+      NOT encodable: nko, tod-oirat (no source PDFs).
 - [x] 06-din-columns.md — DIN 1460-2, 31635 fas/kur/urd, 31636 yid,
       32708 — LANDED: all columns in the corpus (din-fas/kur/urd-31635,
       din-yid-31636, din-jpn-32708; corpus check 2026-11-09)
@@ -59,3 +73,17 @@ every change source-anchored to the reference documents (test-vs-map
 decided per failing row against the source page). This is the corpus's
 first zero-known-failure state.
 - 16-workflow-extension.md — ISC workflow layer: rulesets, override, language conditioning (spec for owner review)
+
+## Site catalogue follow-up (owner-gated dependency)
+
+The site's `src/data/maps-catalogue.json` regeneration against the
+428-map corpus is BLOCKED on an npm release of `interscript-ts`
+carrying the ISC-WF-1 parser: the site pins the published
+`interscript@^5.5.1`, which cannot parse the two WF-1 pilot maps
+(`alalc-tha-Thai-Latn-2011`, `bgnpcgn-tha-Thai-Latn-2002`) — the
+regenerated catalogue would add the 24 batch maps but silently drop
+those two live entries (a site regression), so it was not committed.
+Sequence for the owner: release an npm version with WF-1 (version
+number is the owner's call), then re-run `npm run
+generate:catalogue` on the site and PR the regenerated file
+(428/428, zero drops).