[squirelworks]

Requested by Squirrel · project thread
Scope: languages/greek/, languages/hebrew/, languages/georgian/, languages/armenian/, languages/thai/, languages/registry.js, languages/reference.js

Wave 13, batch 1 (W13.1 to W13.3) for the first five languages the plan names.

Language	grammar.js	vocab.js	reference.json
Greek	6 lessons, 30 drills	stays in vocabulary-learner/decks/ until W11.5 moves it	11 cards
Hebrew	6 lessons, 30 drills	10 decks, 146 words	10 cards
Georgian	6 lessons, 30 drills	10 decks, 145 words	10 cards
Armenian (Eastern)	6 lessons, 30 drills	10 decks, 142 words	10 cards
Thai	6 lessons, 30 drills	10 decks, 139 words	10 cards

• Files follow the W11.1 schema (W11.1: language data schema and loader #100): each meta.json lists them under data, and LSD.load(id, kind) returns them.
• Grammar and vocab use the same shapes as grammar-learner/data/*.js and vocabulary-learner/decks/*.js, so they appear in those learners once W11.4 and W11.5 switch them to LSD.has / LSD.load (agreed with the thread doing the move). No learner code changes here.
• Reference cards show up in the Reference Learner now, next to the Alphabets letter cards.
• Content is written for this project from general knowledge of each language; nothing is copied from paid sources.

Checks

• python3 languages/build_registry.py --check passes.
• Every drill's answer is among its options and no deck repeats a word (local script).
• Smoke on the Desk server (http://127.0.0.1): LSD.load returns grammar, vocab and reference for all five languages, and the Reference Learner shows their cards with no page errors.
• pr_guard.py passes.

---

Generated by Claude Code

[github-actions]

Merge verdict: SAFE TO MERGE (PR guard passed on 581fe68). Use "Squash and merge".

PR guard output

Merge base 88551662db; 21 files changed; 21 files touched by the branch's own commits.

Scope: languages/greek/, languages/hebrew/, languages/georgian/, languages/armenian/, languages/thai/, languages/registry.js, languages/reference.js

PR guard: passed