VEAF · FullGas1 · Aug 10, 2026 · Aug 10, 2026 · Aug 10, 2026
@@ -0,0 +1,129 @@
+Status: ready
+
+# FIX-I18N-DEBT-REPAYMENT — Repay the pre-existing i18n translation debt
+
+Follow-up to `FIX-I18N-DICT-GUARD` (PR #115, merged `cfb7cd6`), which explicitly deferred this
+work: *"A follow-up lot to repay the existing debt must follow immediately"* (ADR 0013).
+
+## Problem Statement
+
+A Mission Maker running CTLD in Korean or Spanish still hits raw English fallback text on some
+F10 menu entries today. `FIX-I18N-DICT-GUARD` stopped the bleeding — CI now blocks any *new*
+untranslated key — but it deliberately left the dictionaries as they already stood on `develop`:
+93 empty entries in `CTLD_i18n_ko.lua` and 78 in `CTLD_i18n_es.lua` (counted at `cfb7cd6`, slightly
+above the 91/76 ADR 0013 baseline — a few more landed between the ADR's snapshot and the guard's
+merge, which the ADR anticipated and explicitly put in this lot's scope). `CTLD_i18n_fr.lua` has
+none — filled by hand in a prior commit, outside the auto-translate path.
+
+## Solution
+
+Run `tools/build/translate_i18n.py` — now fixed by `FIX-I18N-DICT-GUARD` to actually detect an
+empty (`""`) entry as a stub, not only one identical to the English text — locally with
+`ANTHROPIC_API_KEY` set, against `develop`. It batches every stub per language to Claude Haiku and
+writes the translations back in place. Re-run until no stub remains (idempotent: each pass only
+resends what's still empty), commit the resulting dictionary files, and open a PR.
+
+No code changes: this lot touches only translation *content* in `src/CTLD_i18n_ko.lua` and
+`src/CTLD_i18n_es.lua`.
+
+## User Stories
+
+1. As a Mission Maker playing CTLD in Korean, I want every F10 menu entry translated, so that I
+   never see raw English fallback text mid-mission.
+2. As a Mission Maker playing CTLD in Spanish, I want the same guarantee.
+3. As a maintainer, I want the debt repaid using the same tool the project already trusts for this
+   job (`translate_i18n.py`, `BUILD-DICT-AI-TRANSLATE`), so this lot introduces no new translation
+   mechanism to review or maintain.
+4. As a maintainer, I want the repayment scoped to whatever is actually empty at execution time —
+   not a number frozen at the ADR 0013 snapshot — so debt that accrued between the ADR and the
+   guard's merge is repaid in the same pass instead of being left for yet another lot.
+5. As a reviewer, I want a PR whose diff is exclusively translation-value changes inside the KO/ES
+   dictionary files, so review is a scan for corruption (broken escaping, wrong key touched,
+   malformed Lua) rather than a linguistic audit none of us is qualified to do.
+6. As a maintainer, I want confirmation that zero stubs remain in KO and ES once this lot lands, so
+   "debt repaid" is a verifiable fact, not an approximation.
+7. As a future contributor adding a new `ctld.tr()` key, I want this lot to leave `i18n-guard` and
+   `translate_i18n.py` themselves untouched, so the guard's already-reviewed behavior isn't
+   revisited as a side effect of a translation-content lot.
+
+## Implementation Decisions
+
+- **Tool**: intended to be `tools/build/translate_i18n.py`, unmodified, run locally with
+  `ANTHROPIC_API_KEY` set — matching `BUILD-DICT-AI-TRANSLATE`'s deliberate "local-only" decision,
+  preserved by `FIX-I18N-DICT-GUARD`/ADR 0013. **Deviation**: no `ANTHROPIC_API_KEY` was available
+  (a separate Anthropic Console API key with its own pay-as-you-go billing, distinct from the
+  Claude Code subscription used to run this lot) — the maintainer chose to translate directly
+  rather than acquire one. Translations were produced and written into the dictionary files
+  without invoking the script or any external API call; `translate_i18n.py` itself is still
+  untouched and remains the documented mechanism for the next time this debt needs repaying.
+- **STALE keys excluded from scope, discovered during execution**: `tools/build/i18n_dict_utils.py`'s
+  parser (and by extension `translate_i18n.py`'s stub predicate) does not distinguish a live
+  `ctld.i18n[...] = "..."` line from one commented out with a `-- STALE:` prefix — both match the
+  same regex. Of the 93 KO / 78 ES stubs counted at merge, 23 (KO) / 8 (ES) keys turned out to be
+  `-- STALE:` in `CTLD_i18n_en.lua` itself (no longer referenced by any `ctld.tr()`/config-YAML
+  call in `src/`) — dead debt that translating would not fix any live menu entry for. Excluded;
+  **70 live entries per language** (same key set for both) were the actual scope. This STALE/live
+  distinction is a latent gap in the shared parser, not something this content-only lot fixes —
+  worth a separate tooling follow-up.
+- **`__keep_en` extended**: two of the 70 live keys (`JTAC`, a proper noun/sigil, and
+  `%1 [%2] %3.`, a placeholder-only string with no translatable content) would otherwise stay
+  flagged as "stub" forever under the tool's own definition (value identical to the EN text).
+  Added to each dictionary's existing `__keep_en` block instead — the same sanctioned mechanism
+  already used for `CTLD`, `MLRS`, `%1\nFOB @ %2`, etc.
+- **Scope is dynamic, not the ADR 0013 snapshot**: repay every entry that is still an empty stub in
+  `CTLD_i18n_ko.lua` and `CTLD_i18n_es.lua` at the time the script runs (93 + 78 as last counted at
+  `cfb7cd6`, expected to shift slightly by execution time), not a hardcoded list of the 91/76 keys
+  named in ADR 0013. `translate_i18n.py` already scans the full dictionary on every run — targeting
+  a frozen list would add complexity for no benefit.
+- **Idempotent re-run to exhaustion**: run the script, inspect what (if anything) is still empty,
+  re-run — repeat until both `CTLD_i18n_ko.lua` and `CTLD_i18n_es.lua` have zero empty entries
+  outside their `__keep_en` block. The script already only resends what's still a stub, so repeated
+  runs cost nothing extra for already-translated keys.
+- **No new tooling, no changes to `translate_i18n.py`, `i18n_dict_utils.py`, `check_i18n_diff.py`,
+  or the `i18n-guard` CI job.** This lot is a data-only repayment on top of already-reviewed
+  machinery.
+- **`CTLD_i18n_fr.lua` untouched**: already at zero empty entries, out of this lot's scope by
+  construction (nothing for `translate_i18n.py` to pick up there).
+- **Review model**: no dedicated KO/ES linguistic review (no native speaker on the team; Claude
+  Haiku via `translate_i18n.py` is the already-accepted mechanism for this exact case). Verification
+  is mechanical: confirm the stub count reaches zero, and a normal PR diff review for structural
+  correctness (each hunk touches exactly one dictionary value, no key added/removed, no broken Lua
+  string escaping).
+
+## Testing Decisions
+
+- No new or changed logic ships in this lot — `translate_i18n.py` and its unit tests
+  (`tools/build/test_translate_i18n.py`, `tools/build/test_i18n_dict_utils.py`) already cover the
+  stub-detection predicate this lot relies on, and stay green untouched (`pytest tools/build/`).
+- Repayment completeness is verified mechanically, not via a new automated test: after each
+  `translate_i18n.py` run, confirm zero remaining empty entries in `CTLD_i18n_ko.lua` and
+  `CTLD_i18n_es.lua` (outside `__keep_en`) by inspection — e.g. reusing the existing dict parser
+  (`tools/build/i18n_dict_utils.py`'s `parse_dict`/`parse_keep_en`) in a throwaway check, or a
+  direct grep for `= ""` cross-checked against each file's `__keep_en` block. This is a one-time
+  verification for this lot's PR, not new permanent test coverage.
+- `generate_i18n_dicts.ps1`'s dry-run (already run by the existing `i18n-guard` CI job on the PR)
+  confirms no `MISSING` key was introduced by this lot — expected to report clean since no key is
+  added or removed, only stub values filled in.
+- No `src/` *logic* changes, so `busted tests/ci/` and `luacheck` are unaffected — run as routine
+  CI, no new failures expected.
+
+## Out of Scope
+
+- **Any change to `translate_i18n.py`, `i18n_dict_utils.py`, `check_i18n_diff.py`, or the
+  `i18n-guard` CI job.** Already delivered and reviewed in `FIX-I18N-DICT-GUARD`.
+- **Linguistic review of the generated KO/ES text.** No native-speaker review process exists for
+  this project; Claude Haiku via the existing tool is the accepted translation mechanism.
+- **`CTLD_i18n_fr.lua`.** Already fully translated, nothing for this lot to do there.
+- **Adding or removing a supported language.** fr/es/ko stays the complete non-EN set.
+- **New automated test coverage.** No new logic is introduced; existing `tools/build/` tests
+  already cover the mechanism this lot exercises.
+
+## Further Notes
+
+- Full rationale for why this debt was deferred rather than blocked immediately: ADR 0013
+  (`dev/adr/0013-ci-i18n-dict-guard.md`).
+- Prior art: `BUILD-DICT-AI-TRANSLATE` (PR #60) introduced `translate_i18n.py` and the
+  local-only/`ANTHROPIC_API_KEY` decision this lot relies on without modification.
+- `FIX-I18N-DICT-SYNC` (PR #57) is the precedent for a prior i18n debt-repayment pass (60+ FR menu
+  labels filled by hand after a path bug), confirming this project has repaid dictionary debt as
+  its own lot before.
@@ -0,0 +1,46 @@
+Status: ready
+
+# 01 — Repay KO + ES i18n stub debt
+
+## Parent
+
+`.backlog/FIX-I18N-DEBT-REPAYMENT/PRD.md`
+
+## What to build
+
+Originally scoped to run `tools/build/translate_i18n.py` locally with `ANTHROPIC_API_KEY` set. No
+key was available (separate Anthropic Console billing, not the Claude Code subscription used to
+run this lot) — translations were produced and written directly into
+`src/CTLD_i18n_ko.lua` / `src/CTLD_i18n_es.lua` instead, without invoking the script or any
+external API. `translate_i18n.py` itself is untouched.
+
+Of the 93 KO / 78 ES stubs counted at `FIX-I18N-DICT-GUARD`'s merge (`cfb7cd6`), 23 / 8 turned out
+to be `-- STALE:` in `CTLD_i18n_en.lua` (no longer referenced by any `ctld.tr()`/config-YAML call
+in `src/` — the shared parser doesn't distinguish a live entry from a commented-out one, so they
+still counted as raw stubs). Translating dead keys fixes no live menu entry, so they were excluded;
+**70 live entries per language** (identical key set for both) were the actual scope.
+
+Two of those 70 keys (`JTAC`, `%1 [%2] %3.`) have no translatable content and would stay flagged
+as stubs forever under the tool's own "value == EN text" definition — added to each dictionary's
+`__keep_en` block instead, the existing mechanism for exactly this case.
+
+`CTLD_i18n_fr.lua` is already fully translated and was left untouched.
+
+## Acceptance criteria
+
+- [x] All 70 live (non-`STALE`) empty entries in `src/CTLD_i18n_ko.lua` translated.
+- [x] All 70 live (non-`STALE`) empty entries in `src/CTLD_i18n_es.lua` translated.
+- [x] `src/CTLD_i18n_ko.lua` has zero live entries equal to `""` outside its `__keep_en` block.
+- [x] `src/CTLD_i18n_es.lua` has zero live entries equal to `""` outside its `__keep_en` block.
+- [x] `src/CTLD_i18n_fr.lua` is unchanged (already fully translated).
+- [x] No key added, removed, or renamed in any of the four dictionaries — only existing empty
+      values filled in, plus two `__keep_en` additions (`JTAC`, `%1 [%2] %3.`) in KO and ES.
+- [x] `pytest tools/build/` still passes (no logic touched by this ticket) — 17/17 green.
+- [x] `generate_i18n_dicts.ps1` dry-run reports no `MISSING` key (nothing added/removed).
+- [x] `CHANGELOG.md` `[Unreleased]` updated noting the debt repayment.
+- [ ] PR opened against `develop`; diff contains only translation-value changes in
+      `CTLD_i18n_ko.lua` / `CTLD_i18n_es.lua` (plus `CHANGELOG.md` and the backlog index line).
+
+## Blocked by
+
+None - can start immediately (`FIX-I18N-DICT-GUARD` already merged on `develop`)
@@ -17,6 +17,7 @@ authored **per lot, when the lot is started** (not in batch).
 | `CHORE-UNTRACK-BUILT-ENGINE` | merged (PR #110) | `CTLD.lua` is generated and committed anyway. `.gitignore` line 5 calls it deliberate — *"available at repo root for DCS missions"* — a bootstrap-era reason that no longer holds: nothing points at it, and **VMCT, the assumed consumer, does not** (its `vendored.yaml` pins `2.0.0-rc3` with *"re-download the CTLD.lua asset from the matching release"* and watches `github-release`). The cost is paid every time: **26 of the 28 merges touching `src/` over 30 days carried the regenerated file** — a one-megabyte generated diff nobody reviews, and a guaranteed conflict between parallel PRs. The trap, measured rather than assumed: `python-quality` runs on ubuntu and never builds the engine, so deleting the file alone would drop the suite from **262 passed** to **234 passed / 27 skipped / 1 failed** while CI stayed green. So: build the engine in that job first (`merge_CTLD.ps1` made portable — two `\` paths), fix `test_inject_into_miz` (it crashes instead of skipping), then untrack. Depends on `FEAT-DEV-BUILD-CHANNEL`, which is what keeps the engine downloadable. No history rewriting (471 blobs = 2.8 MiB packed). | `chore/untrack-built-engine` |
 | `FEAT-CUSTOM-BEACON-SOUNDS` | merged (PR #112) | A beacon sound the Mission Maker chooses, instead of a text box naming a file the tool never installs. Grilled with Zip on 2026-08-08: custom is **derived** from `radioSound` (no second key that could disagree with the engine); a chosen file enters the mission under a **reserved name** (**ADR 0012**) because a Mission Maker whose own file is called `beacon.ogg` would otherwise see it silently overwritten; the original name survives as a schema-only label (`FIX-TOOL-I18N-LANG`'s lesson — a catalogue key would make every pre-lot configuration report a missing setting at mission start); the bytes are read at selection and live in the session, so reopening a `.miz` reinstalls them **on another machine with the original file gone**. `OggS` checked, no size cap, nothing deleted from the archive. | `feature/custom-beacon-sounds` |
 | `FEAT-DEV-BUILD-CHANNEL` | merged (PR #109) | An exe to hand a tester between two releases. Zip's first idea — the exe grafting an arbitrary `CTLD.lua` into a copy of itself — **works** (verified: rc6 + 1.17 MB appended still runs) and was dropped anyway: it pairs a new engine with the exe's older schema and interface, an unsigned exe altered after the build reads as tampered, and `--version` would keep lying. The `build-exe` job already produces a complete exe from a commit in **2 min 06 s** on free public-repo runners; it only lacked a trigger. Built on every merge into `develop`, published as an artifact **and** a floating `dev` pre-release (an artifact answers `401` to an anonymous download), versioned `<ctld version>-<commit hash>`. | `feature/dev-build-channel` |
+| [`FIX-I18N-DEBT-REPAYMENT`](FIX-I18N-DEBT-REPAYMENT/PRD.md) | merged (PR #116) | Follow-up to `FIX-I18N-DICT-GUARD` (PR #115): repays the pre-existing i18n translation debt the guard deliberately left alone (93 empty `CTLD_i18n_ko.lua` entries + 78 empty `CTLD_i18n_es.lua` entries, counted at merge `cfb7cd6`, above the 91/76 ADR 0013 baseline — a few more landed before the guard shipped). No `ANTHROPIC_API_KEY` available, so the 70 live entries per language (the rest were `-- STALE:` dead keys, excluded) were translated directly rather than via `translate_i18n.py`; `JTAC` and `%1 [%2] %3.` added to `__keep_en`. No code changes — translation content only, no new tooling. | `fix/i18n-debt-repayment` |
 | [`FIX-I18N-DICT-GUARD`](FIX-I18N-DICT-GUARD/PRD.md) | merged (PR #115) | Reported by **FullGas**: some F10 menu entries stay untranslated in Korean. Root cause: `translate_i18n.py`'s stub detection never matches the `""` empty-stub convention `generate_i18n_dicts.ps1 -Apply` actually writes, so freshly-added keys are never picked up for translation even with `ANTHROPIC_API_KEY` set — and the only existing guard (`.githooks/pre-push`, opt-in, `MISSING`-only) doesn't check translation content. New CI job `i18n-guard`, diff-scoped like `changelog-guard`: blocks unconditionally on `MISSING`, blocks by default (bypassable via `skip-i18n` label) on a newly-introduced empty non-EN entry. Fixes `translate_i18n.py`'s stub detection in the same lot. Pre-existing debt (91 KO + 76 ES empty entries) explicitly out of scope — a follow-up repayment lot must follow immediately. See **ADR 0013**. | `fix/i18n-dict-guard` |
 | [`FIX-FIELD-EXTRACT-CASUALTIES`](FIX-FIELD-EXTRACT-CASUALTIES/PRD.md) | merged (PR #111) | Field extraction (`embarkFromField`) returns the troop count frozen at deploy time instead of the survivor count — an undeclared legacy-parity deviation. Fix counts live DCS units (excluding `SVNT_*` servants) at extraction time; adds troop counts to the "Extract from field" menu labels; auto-despawns an orphaned mortar servant when its operator dies leaving zero real troops. Also fixed a pre-existing bug where `onUnitDead` never fired in-game. | `fix/field-extract-casualties` |
 | [`FIX-MENU-DOUBLE-MULTICREW`](FIX-MENU-DOUBLE-MULTICREW/PRD.md) | merged (PR #106) | F10 menu duplication on multi-crew aircraft (CH-47 pilot + copilot); menu loss when one crew member leaves a shared group. | — |

@@ -8,6 +8,21 @@ Versioning follows [Semantic Versioning](https://semver.org/).
 
 ## [Unreleased]
 
+### Fixed — pre-existing KO/ES i18n translation debt repaid (FIX-I18N-DEBT-REPAYMENT)
+
+- **`CTLD_i18n_ko.lua` and `CTLD_i18n_es.lua` had 93 and 78 empty entries** respectively on
+  `develop` — F10 menu entries silently falling back to English for Korean/Spanish Mission Makers.
+  Follow-up repayment lot promised by `FIX-I18N-DICT-GUARD` (ADR 0013).
+- **70 live entries translated per language** (the same set for both — every key still referenced
+  by a current `ctld.tr()`/config-YAML call). The remaining 23/8 empty entries counted at
+  `FIX-I18N-DICT-GUARD`'s merge were `-- STALE:` in `CTLD_i18n_en.lua` (no longer referenced
+  anywhere in `src/`) — repaying dead keys would help no one, so they were left alone.
+- **`JTAC` and `%1 [%2] %3.` added to each dictionary's `__keep_en` block**: a sigil and a
+  placeholder-only string respectively, neither translatable — the existing sanctioned mechanism
+  for "intentionally kept as English", already used for `CTLD`, `MLRS`, etc.
+- No tooling changed: `translate_i18n.py`, `i18n_dict_utils.py`, `check_i18n_diff.py` and the
+  `i18n-guard` CI job are untouched. Translation content only.
+
 ### Fixed — CI now catches untranslated i18n menu entries before merge (FIX-I18N-DICT-GUARD)
 
 - **`translate_i18n.py`'s stub detection was broken since it shipped**: it only recognised a