Sweep: link fixes, splash i18n parity, neuronetz badge, mobile, training source

LINK AUDIT — every internal link now resolves to a real page:
  - BrandHeader doc-nav 'Showcase' target → /{locale}/showcase/projects/
    (the showcase/ directory has no index, only patterns/projects)
  - Translated-slug bugs across de/es/fr/ja: ~30 broken links from the
    overnight translator that auto-localised path segments. Mapped back to
    the actual English slugs:
      de:  warum-nibiru → why-nibiru, kuenstliche-intelligenz/orakel → ai/oracle,
           kI/modul/* → ai/module/*, praesentation/projekte → showcase/projects
      es:  ai/modulo/* → ai/module/*, diseno/* → design/*,
           porque-nibiru / por-que-nibiru → why-nibiru,
           presentacion/proyectos → showcase/projects
      fr:  ia/module/* → ai/module/*, ia/module/formation → ai/module/training,
           pourquoi-nibiru → why-nibiru, presentation/projets → showcase/projects
      ja:  ai/milestones → ai/roadmap
  - Cross-locale leaks: bare `/ai/oracle` (no locale) → `/{locale}/ai/oracle/`
    in de/index.mdx, fr/index.mdx, fr/ai/module/overview.md, en/ai/corpus.md
  - `/en/start/` (which 404s — start/ has no index) was hardcoded in five
    design/components.md atelier-button hrefs across all locales →
    `/{locale}/start/installation/`
  - `/en/reference` removed from en/downloads.mdx — the reference doc tree
    isn't built yet; replaced with a github link to the v2 markdown source
  - Collapsed stray double-slashes (`/de/why-nibiru//` etc.) introduced by
    the slug-replacement sed sweep

Final audit shows 0 broken internal links (down from ~37).

TRAINING SOURCE NOW SHIPS — root cause of "I cannot find the training data
in the repository": the curated source files were gitignored.
  - .gitignore at scripts/extraction/ now whitelists framework-reference-v2.md
    and lora-augmentation.summary.txt alongside lora-augmentation.jsonl
  - The 1620-line v2 reference, the 323-record augmentation jsonl, and the
    summary report all enter the repo so the production Docker build sees
    them and contributors can find them by browsing gitea

NEURONETZ AI DEEPLINK BADGE — small "AI by Neuronetz ↗" pill in the splash
footer's bottom strip. Logo mark mirrored locally to
public/img/external/neuronetz-mark.svg (pulled from neuronetz.ai/favicon.svg)
so the page doesn't hot-link off-domain on every paint. Magenta border on
hover; opens neuronetz.ai in a new tab with rel=noopener.

SPLASH I18N PARITY — de/es/fr/ja index.mdx now import + render the same
component stack as en (CometTrail · MmvcStage · MissionControl · LaunchSequence
· SpacecraftGrid · EditorialContent · LandingFooter · ToTop · LandingScripts),
so every locale shows the full splash structure. The component bodies
themselves are still English (proper i18n is the next step); for now this
brings structural parity.

MOBILE RESPONSIVE SWEEP:
  - LandingFooter: 4-col grid stacks 2-col @ 768px and 1-col @ 480px;
    bottom strip wraps vertical at 480px
  - MmvcStage: 5-step progress rail tightens its gaps under 720px and
    drops the bar segments entirely under 480px so the labels fit
  - Docs bridge §11: tighter H1/H2 spacing, breadcrumbs/doc-meta on
    narrow viewports, pagination cards stack 1-col, help-strip stacks
    vertical, tables get horizontal-scroll on overflow
  - Doc-header: nav-version chip hides under 480px so the search-pill
    + brand fit comfortably

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

docs/scripts/extraction/.gitignore

@@ -1,9 +1,13 @@
 # Extraction outputs from research-agent runs are typically large + transient.
-# Keep them local by default — except the curated augmentation file which
-# needs to ship for the Docker build to fold it into the LoRA corpus.
+# Keep them local by default — except the curated files that need to ship so
+# the Docker build can fold them into the LoRA corpus AND so contributors
+# can find the source-of-truth reference in the repo.
 *.md
 *.json
 *.jsonl
+*.txt
 
-# …with one explicit exception:
+# …with explicit exceptions:
+!framework-reference-v2.md
 !lora-augmentation.jsonl
+!lora-augmentation.summary.txt

docs/scripts/extraction/lora-augmentation.summary.txt

@@ -0,0 +1,68 @@
+Nibiru Framework — LoRA Augmentation Corpus Summary
+=====================================================
+
+Source:    /home/stephan/PhpstormProjects/Nibiru/docs/scripts/extraction/framework-reference-v2.md
+Output:    /home/stephan/PhpstormProjects/Nibiru/docs/scripts/extraction/lora-augmentation.jsonl
+Format:    JSONL (one record per line)
+Records:   323
+
+Validation
+----------
+All 323 lines are valid JSON. All records carry the required keys:
+instruction, input, output, metadata.{language, source, section, kind, citations}.
+(Verified with `jq` schema check — 0 records failed.)
+
+Breakdown by `kind`
+-------------------
+code-recall   121
+workflow       97
+inikey         29
+comparison     26
+gotcha         16
+debug          16
+edge-case      15
+refactor        3
+
+Coverage
+--------
+All 14 top-level sections of the v2 reference are exercised:
+- §0  Executive Summary
+- §1  Namespace Map
+- §2  Singletons & Factories (Config, Pdo, Db, Form, Controller, View)
+- §3  Database Layer (drivers, autogenerator pipeline, multi-database)
+- §4  Module System (structure, registry, loading flow, plugin pattern, lifecycle)
+- §5  Form System (architecture, substitution, type classes, ordering, div wrap, assembly, production usage)
+- §6  Request/Response Lifecycle
+- §7  INI Key Reference
+- §8  Namespace Conventions
+- §9  Constants
+- §10 Production Patterns NOT in Core
+- §11 Gotchas & Footguns
+- §12 Request Handling Trace
+- Appendices: File Index, Quick Reference
+
+Cross-section workflow / debug / refactor records (~30) cover full-stack scenarios:
+adding a new module, adding a new table, switching DB sections per model,
+debugging Postgres-mode mask footgun, end-to-end form-to-DB save flow, etc.
+
+Sample records (3 for sanity-check)
+-----------------------------------
+
+Sample 1 (kind=code-recall, line 14):
+{"instruction":"Show me the exact `__construct()` body the autogenerator produces for a MySQL model.","input":"","output":"```php\npublic function __construct()\n{\n    Pdo::settingsSection('DATABASE');\n    self::initTable( self::TABLE );\n}\n```\nFrom `loach.mssql.database-connection/application/model/WarehouseLoach/timeanddate.php:30-34`. The first line writes the static `Pdo::$section`; the second registers the table schema with the adapter's static `$table` array.","metadata":{"language":"en","source":"framework-reference-v2","section":"2.2 Pdo","kind":"code-recall","citations":["loach.mssql.database-connection/application/model/WarehouseLoach/timeanddate.php:30-34","framework-reference-v2.md §2.2"]}}
+
+Sample 2 (kind=inikey, line 100):
+{"instruction":"Which INI key sets the PHP timezone?","input":"","output":"`[SETTINGS][timezone]`. Read by `Dispatcher::run()`. Defaults to `UTC`. From `framework-reference-v2.md §7 [SETTINGS]`.","metadata":{"language":"en","source":"framework-reference-v2","section":"7 INI [SETTINGS]","kind":"inikey","citations":["framework-reference-v2.md §7"]}}
+
+Sample 3 (kind=workflow, line 220):
+{"instruction":"How does the autogenerator handle a brand-new application with zero existing model files?","input":"","output":"`Model::createOutFolder()` at `core/c/model.php:28` does `mkdir -p` (chmod 0777) for the target folder, then `createClassFiles()` at line 37 loops through scanned tables (or one table if `--table=` was passed) and calls `generateClassByTableName($table)` for each. For each class file:\n1. If overwrite mode, `unlink` existing.\n2. Otherwise `fopen('w'); fclose()` to create an empty stub.\n3. Build template, do all 8 placeholder substitutions plus driver/odbc switching.\n4. Write file and chmod 0777.\n\nSo zero existing files becomes a full set of generated models on the first request. From `framework-reference-v2.md §3.2`.","metadata":{"language":"en","source":"framework-reference-v2","section":"3.2 Autogenerator Pipeline","kind":"workflow","citations":["core/c/model.php:28","core/c/model.php:37","core/c/model.php:53","framework-reference-v2.md §3.2"]}}
+
+Notes for downstream merge
+--------------------------
+- Every record has a real file:line citation OR a quoted code block.
+- Code blocks are quoted from the v2 reference; nothing is invented.
+- Phrasing varies across natural developer styles ("Walk me through…",
+  "Quick question:…", "Why does…", "I'm trying to…", "Show me…",
+  "I'm getting `Class … not found`…", etc.).
+- The `kind` taxonomy follows the spec: edge-case, comparison, workflow,
+  debug, refactor, code-recall, inikey, gotcha.