Commit graph

2306 commits

Author SHA1 Message Date
Till JS
ad5987f1dd feat(visibility): M6 soft-migrate isPublic→visibility on memoro/cards/presi/uload
Brings four legacy isPublic-only modules onto the unified visibility
system. The Picker (private/space/public) replaces the ad-hoc
boolean toggle in memoro + cards + presi DetailViews. Each store
now mirrors `visibility ↔ isPublic` on every flip so older readers
(search index, snapshot pipelines, sync server) keep working
through the soak window — M6.1 will hard-drop the legacy field
once the field has propagated to all rows.

Per-module:
- memoro: setVisibility on memosStore + Picker in DetailView
  properties row. Picker reads the unified `visibility` with a
  fallback to legacy isPublic.
- cards (decks): replaces the "Öffentlich Ja/Nein" toggle button
  with the Picker. createDeck initializes both fields; updateDeck
  mirrors when callers pass legacy isPublic.
- presi (decks): same pattern as cards. setVisibility added to the
  store factory's exported surface.
- uload (tags): no active CRUD UI, so this is type-only soft-migrate
  + seed-data update. Future tag-management view writes visibility
  directly. No store mutation method needed yet.

Out of scope (intentional):
- picture isPublic hard-drop: deferred. Picture has been on
  visibility since M3 (commit 0e9f574df), the soak window is mature
  enough to consider, but pre-launch there's no urgency. Defer to
  M6.1 with the rest of the legacy-field cleanup.
- events isPublished/publicToken: STAYS. This isn't legacy — it's
  the orthogonal mana-events RSVP-snapshot system. visibility
  controls website-embed eligibility; isPublished controls RSVP
  page existence. Different concerns; both stay.

Picker hides 'unlisted' for all four (no server-publish-snapshot
flow wired for these collections — same pattern as habits/quiz/events).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 14:11:07 +02:00
Till JS
bd559e739e polish(picture): clean borderless lightbox — image-first, meta in the corner
Feedback on the previous lightbox: the image was capped at 60vh
inside a bordered card, with metadata + action buttons stacked
underneath in their own panel. Result: ~40% of the modal surface
was chrome around a small picture, and the eye landed on the card
edge before the actual generation. The user asked for image-first:
fill the height, no module border, gray text in the bottom-right
corner.

New layout:

- Borderless backdrop. The card frame is gone; the image sits
  directly on a near-black overlay (rgba(0,0,0,0.92)) so nothing
  competes for the eye.
- Image fills the available area via `max-h-full max-w-full
  object-contain` inside a flex-center wrapper, with 6/10-rem
  outer padding so the picture doesn't kiss the viewport edges
  on small screens.
- Metadata moves to a small grey overlay in the bottom-right
  corner — prompt on top, then a single inline detail line
  (model · dimensions · date) separated by middle dots. Right-
  aligned so longer prompts wrap toward the image edge instead
  of the centre.
- Close becomes a circular icon-button in the top-right (X), no
  longer a footer button. ESC + backdrop-click still close.
- Caller-supplied actions (the `actions` snippet) move to the
  bottom-left so they don't fight the meta block visually.

Colour treatment uses literal white-alpha + black-alpha values
in a scoped `<style>` block instead of theme tokens. The lightbox
always sits on a literal near-black backdrop regardless of which
theme is active, so theme-aware muted tokens would render too
dark in light themes. The validator's brand-literal escape hatch
(see scripts/validate-theme-utilities.mjs comment) covers this
exact case.

Empty-state (publicUrl missing) gets a small SquaresFour icon in
a soft white-alpha tile so the modal still has a visible centre
when an image fails to load.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 14:10:11 +02:00
Till JS
d62ae8f1e3 fix(workbench): dedup duplicate Home scenes accumulated by seeding race
The Home-seeder in workbench-scenes.svelte.ts writes new scenes without
spaceId, so the creating-hook stamps them with the _personal:<userId>
sentinel. The per-space dedup check filters by the real space UUID and
never finds them — every login adds another Home row, and every visit
to a non-personal Space (Brand/Family/Team) drops yet another seed
into the personal Space.

This is Schicht D-soft of the broader cleanup plan
(docs/plans/workbench-seeding-cleanup.md): a one-shot dedup pass that
collapses duplicate "Home" rows per spaceId, merging openApps from the
losers into the survivor (most apps wins, ties by most-recent
updatedAt) and soft-deleting the rest so mana-sync propagates the
cleanup to other devices. Touches only rows that look like fresh
default seeds — anything customized (description, wallpaper, agent
binding, scope tags, non-Home name) is left alone.

Wired in two places: a Dexie v48 upgrade so it runs once per device on
schema bump, and a belt-and-suspenders pass in (app)/+layout.svelte
right after reconcileSentinels() to catch the edge case where
sentinel-stamped rows just collapsed into the same UUID group as
already-reconciled rows.

The structural fix that prevents new duplicates from ever forming
(per-space-seeds registry + deterministic seed ids +
creating-hook hardening) ships in follow-up commits per the plan.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 14:08:32 +02:00
Till JS
ac12b61de2 fix(writing): replace emojis with Phosphor icons in workbench empty-state
Visual cleanup for the workbench-card hero. Emojis were dropping the
design fidelity vs the rest of the Mana UI, which uses Phosphor icons
exclusively in chrome surfaces.

- Action bar 🎨 → <Palette /> in a 2x2-rem ghost button (square, no
  text label) so the visual weight matches the primary "+ Neuer Draft"
  button instead of competing with it.
- Hero icon 📝 → <NotePencil weight="duotone" /> in a 3rem rounded
  tinted-sky pill — gives the empty-state a consistent on-brand
  centerpiece without the multi-color emoji look.
- Hero meta line restructured: was a single -prefixed sentence that
  wrapped awkwardly on narrow workbench cards. Now a centered <ul> of
  short uppercase chips ("12 TEXTARTEN · 9 STILE · 7 QUELLEN · E2E-
  VERSCHLÜSSELT") with a Sparkle icon only on the first item.
- Quick-start tiles 📝/📄/✉️/💬/💌/🎤 → kind-specific Phosphor icons
  via a QUICK_ICON map (blog→Newspaper, essay→Notebook, email→Envelope,
  social→ChatCircle, letter→Envelope, speech→Microphone). Icons inherit
  the muted text colour, switch to sky on hover.
- Hero spacing tightened: padding/margins reduced so the layout breathes
  on a 280–320px-wide workbench card instead of cramming. Quick-grid
  caps at 360px max-width and falls back to 2 cols below 360px.

KIND_LABELS still carries `.emoji` for the populated-state KindTabs and
DraftCard — those stay emoji-flavoured for now since they're inline
text. Icon migration there is a separate sweep when the time comes.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 14:05:26 +02:00
Till JS
59b147f5ee feat(visibility): embed resolvers for habits/quiz/social-events + inspector refresh
Closes the loop on the M5-Rest visibility rollout — flipping a habit,
quiz, or social-event to 'public' now actually surfaces it on the
owner's website embed.

EmbedSourceSchema gains three new sources:
- habits.habits — build-in-public widget. Title + "🔥 N Tage Streak ·
  gesamt M ×". Per-log timestamps + notes stay private (sleep/intake
  patterns are not for public consumption).
- quiz.quizzes — shareable-quiz teaser. Title + "N Fragen · {category}".
  Questions, options, explanations, attempts/scores all stay private —
  the actual play-experience is reserved for a future unlisted-share
  flow.
- events.socialEvents — RSVP-event teaser. Title + formatted start
  date + location + cover image. Hard-gated on the unified `visibility`
  only; the legacy `isPublished` flag is intentionally bypassed so the
  new Picker is the single source of truth (M6 will drop isPublished).

ModuleEmbedInspectorFallback now lists all 12 sources — was only
exposing 2 of the 9 already-wired ones (latent debt unblocking the
new sources from being addable in the editor).

Note: 7 unrelated svelte-check errors exist in
data/scope/dedup-workbench-scenes.test.ts from a parallel session
(spaceId not on LocalWorkbenchScene). Not introduced here.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:58:43 +02:00
Till JS
21dbce6631 feat(writing): smarter empty-state + help-content + de-emphasized Stile link
Three workbench-card UX fixes for the Schreiben module.

Smart empty-state (replaces the small "Noch keine Drafts" line):
- Hero block with icon + 1-line pitch + meta-line (" 12 Textarten · 9
  Stile · 7 Quellen-Typen · End-to-End-verschlüsselt"). Sells the
  module's surface area before the user has anything saved.
- Six quick-start tiles (Blog · Essay · E-Mail · Social · Brief · Rede)
  in a 3-col grid. Tile click → BriefingForm opens with that kind
  pre-selected, skipping the Textart-dropdown for the most common cases.
  Full 12-kind picker remains one click away inside the form.
- Search + KindTabs + status-filter + favorites-toggle now hide when
  drafts.length === 0 — filtering nothing was visual noise. They reappear
  the moment the first draft lands.

De-emphasized Stile link:
- Was a labeled "🎨 Stile" pill competing with "+ Neuer Draft" for
  attention. Now a compact icon-only ghost button (just 🎨) with an
  aria-label so screen readers still get it. Frees the action bar so
  the primary CTA stands alone.

Module help (renders behind the ?-icon in the Workbench shell):
- New `writing` entry in MODULE_HELP with description + 9 features
  (kinds / styles / references / refinement / versioning / visibility /
  export / persona-linkage / token cost) + 5 tips covering keyboard
  shortcuts, -button, the Quellen-flow, drag-source, and custom
  styles. Uses the existing auto-attach via app-registry's onMissing-
  help fallback — no AppPage changes needed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:57:08 +02:00
Till JS
95bedf4625 fix(comic): proper input focus + sichtbare Entfern-Affordance
Drei Tweaks an der Create-Seite, ausgelöst von User-Feedback nach
dem ersten Story-Anlegen:

1. **Inputs/Textarea ohne Focus-Ring.** Title-Input und
   Story-Kontext-Textarea hatten nur `focus:border-primary`
   und sahen im aktiven Zustand fast nicht anders aus als
   inaktiv. Wardrobe's GarmentForm nutzt `focus:ring-1
   focus:ring-primary` plus `disabled:opacity-50` — übernommen.
   Textarea zusätzlich `resize-none` (vertikales Resize von Hand
   ist nett, aber kollidiert mit Padding bei kleiner Karten-Breite).

2. **Face-Tile war ein nackter `<img>`.** Kein border-2, kein
   "Pflicht"-Hinweis — User dachte er könnte's entfernen und
   suchte nach dem X. Jetzt: border-2 border-primary/40 wie
   Body-Tile, plus ein "PFLICHT"-Badge mit Gradient-Overlay am
   unteren Rand. Title="Face-Ref ist Pflicht — kann nicht
   entfernt werden". Damit ist klar: Face = locked.

3. **Body und Garments waren entfernbar, aber das war unsichtbar.**
   - Body-Toggle: bisher gar kein Hover-Feedback im aktiven
     Zustand (Plus-Overlay nur bei inactive). Jetzt im aktiven
     Zustand auf Hover ein rotes X-Overlay über dem Bild
     (group-hover-Pattern, error/60-bg, opacity 0→100).
     Title-Tooltips nochmal verschärft: "Klick zum Entfernen"
     vs. "Klick zum Hinzufügen".
   - Garment-Tiles: das X-Button war 5x5 (20px), abgerundet,
     bg-background/80 — verschwand visuell auf manchen
     Garment-Fotos. Jetzt:
     • Die ganze Tile ist klickbar (Touch-friendly), mit
       hover:border-error/60 (visueller "achtung, klicken
       entfernt das")
     • X-Badge h-6 w-6 (24px) mit Border-Ring, weißer bg-Pille,
       wechselt bei group-hover auf error-bg + weiß. Immer
       sichtbar, nicht erst auf Hover.
   - Heading-Subline ergänzt um "klicke ein Bild oder das ✕,
     um es wieder zu entfernen" — explizite Anleitung.

Comic-Files type-checken sauber.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:51:05 +02:00
Till JS
e0c0791bb5 feat(visibility): pilot extended to habits + quiz + events (M5 finish)
Extends the unified visibility system (@mana/shared-privacy) to three
of the four remaining open modules from the M5 rollout list. Each
module now exposes a private/space/public picker — `unlisted` is
hidden via `disabledLevels={['unlisted']}` because none of the three
have a server-publish-snapshot path yet (M8 territory).

Per-module:
- habits: visibility on LocalHabit + Habit; defaultVisibilityFor on
  createHabit; setVisibility emits VisibilityChanged. Picker in
  HabitDetail right under the header.
- quiz: same pattern on LocalQuiz + Quiz; Picker in EditView meta
  section so quiz-authors flip visibility while editing metadata.
- events (socialEvents): visibility coexists with the legacy
  `isPublished` + `publicToken` flags until M6 consolidation. The
  Picker writes the unified field; publish/unpublish still drives the
  RSVP snapshot. Picker as its own section above RSVPs.

Invoices skipped — `invoiceClients` has no write path yet (the
ClientPicker only reads), and the Invoice document itself is too
sensitive to ever go public. Will land alongside the future
client-portal feature.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:46:50 +02:00
Till JS
d880e89204 feat(writing): rename module display name "Writing" → "Schreiben"
User-facing label only — keeps the route /writing, the module id
'writing', the appId 'writing', and the table prefix writingDrafts/
writingDraftVersions/etc. Just renames the display name in:
- shared-branding/mana-apps.ts (AppSlider label)
- app-registry/apps.ts (Workbench card label)
- the three writing route <title> tags (page tab in browser)

The English code identity stays; the German UI label gets a German
name consistent with Bibliothek / Kontakte / Kalender / Notizen.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:45:02 +02:00
Till JS
32147194fd fix(comic): DataCloneError beim Anlegen + stärkere Hover-States
Zwei Bugs beim ersten Anlegen einer Comic-Story:

1. **DataCloneError "[object Array] could not be cloned"** beim
   `comicStoriesTable.add(newLocal)`. Ursache: StoryForm deklariert
   `characterMediaIds`/`tags` als `$state<string[]>([])` und reicht
   die Proxies direkt an comicStoriesStore.createStory() durch.
   Dexie/IndexedDB's structured-clone refuseiert Svelte-5-State-
   Proxies — `tags`+`panelMeta` werden vorher von encryptRecord zu
   Ciphertext-Strings gewandelt, aber `characterMediaIds` (und
   `panelImageIds`/`tags` falls plaintext) bleiben Arrays und die
   schmieren als Proxy in den IDB-Write rein.

   Fix: Arrays beim Store-Eintritt mit `[...arr]` snapshotten.
   Greift jetzt auch in updateStory() für die gleichen Felder
   (zukünftiges StoryForm-Edit-Mode wäre sonst denselben Bug
   wert). Wardrobe hat das Problem latent auch, fix dort folgt
   wenn jemand auf gleiche Stelle stößt — die Comics-Lösung
   isoliert.

2. **Kein Button-Feel auf der Create-Seite.** StylePicker und
   CharacterPicker hatten zu schwache Tailwind-Hover-Klassen
   (`hover:bg-muted` only) — User las die Tiles nicht klar als
   klickbar.

   StylePicker komplett auf scoped CSS umgezogen (Pattern wie
   PanelModelPicker / wardrobe TryOnModelPicker): hover ändert
   border-color (border-primary/50) UND background (primary/5)
   UND fügt einen schwachen Schatten hinzu, plus active:translate-y-px
   für Touch-Feedback. Active-Tile bekommt klaren primary-Border +
   primary/8-bg + Schatten. Focus-visible mit Outline für
   Keyboard-Nav. role="radiogroup" + aria-checked statt aria-pressed
   für korrekte Semantik.

   CharacterPicker: Body-Ref-Toggle bekommt jetzt hover:border-primary/50
   + hover:shadow-sm und beim aktiven State einen leichten primary-
   Schatten. Add-Garment-Button kriegt hover:border-primary/50 +
   hover:bg-primary/5 + hover:shadow-sm.

5 Encryption-Tests weiter grün. check für comic-Files clean.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 13:39:47 +02:00
Till JS
75c366bff4 test(writing): unit tests for prompt-builder + reference-resolver
64 new tests across two pure-logic surfaces — no Dexie / network /
component setup, runs in <150ms. Plus the LOCAL TIER PATCH revert
that's been waiting for the release window.

prompt-builder.test.ts (39 tests):
- buildDraftPrompt: ghostwriter system + topic/length/kind plumbing,
  optional audience/tone/extra-instructions, preset style injection,
  resolved-references rendering with singular/plural Quelle wording
  and proper bookend markers.
- All five selection prompts (shorten 50–60% / expand 150–180% / tone
  with target / rewrite with instruction / translate with target lang).
- buildTitleSuggestionPrompt: 4–8-word ask, no quotes, no period, no
  prefix; with/without excerpt block.
- cleanSuggestedTitle: now iterative-until-stable so combined artefacts
  ("Titel: \"Hello World\".") collapse in one call. Quote variants
  (straight, curly, German „, French «, single ‚) all stripped via
  asymmetric open/close sets.
- estimateMaxTokens: clamping to [256, 8000], words/chars/minutes
  conversions, fallback when targetLength is null.

reference-resolver.test.ts (25 tests):
- Per-kind shaping for article (siteName-prefix, content/excerpt
  fallback, truncation marker), note (untitled fallback), library
  (book metadata in the label), url (no fetch), kontext (singleton
  via scopedForModule, deletedAt skip), goal (plaintext, no decrypt
  call asserted), me-image (label + tags descriptor, kind fallback).
- Aggregate-budget enforcement in resolveReferences: drops nulls,
  stops adding once MAX_TOTAL_REFERENCE_CHARS is exceeded, but always
  keeps the first ref even if it alone busts the cap (so a single
  large reference doesn't silently produce zero output).

Side-fix: resolver uses `||` for the article content/excerpt fallback
so empty-string content (extraction failures) falls through to the
excerpt — `??` was passing empty strings as valid.

LOCAL TIER PATCH revert: requiredTier flips from 'guest' to 'beta'
in shared-branding/mana-apps.ts. Writing now gates correctly on
release; the comment marker is removed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:57:24 +02:00
Till JS
2e9ec76d60 feat(writing): token-usage in version history + draft drag-source
Two small UX wins.

Version-history shows generation cost
- VersionHistory takes a generations[] prop (DetailView already pulls
  one via useGenerationsForDraft) and looks up each AI version's linked
  Generation by id. When found, renders a monospace cost line under the
  version's wordcount: "1234 → 567 Tokens · 1.4s · ollama/gemma3:4b".
- Skips silently when the generation row isn't there (e.g. older drafts
  before the field was tracked, or a generation that was reverted).
- Lets the user see what each draft cost without digging into the
  Workbench audit timeline.

Drafts as drag source
- DraftCard wires `use:dragSource` with type='draft' + a payload
  carrying id / title / kind / content / wordCount / topic. Cards in
  the Writing list view are now drag origins for any drop target that
  declares acceptsDropFrom: ['draft'].
- App-registry entry for 'writing' gets the matching collection /
  paramKey / dragType / getDisplayData fields so the workbench layer
  treats drafts as full first-class drag-citizens (sibling navigation,
  display fallbacks).
- @mana/shared-ui DragType union extended with 'draft'.

No drop-target wiring yet — articles' acceptsDropFrom can pick up
'draft' as a follow-up, but the M10 ExportMenu's "Als Artikel
speichern" already covers that flow from the editor side.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:31:29 +02:00
Till JS
b7a54ccd10 feat(unlisted-sharing): QR code + per-link expiry picker (M8.5)
SharedLinkControls now renders a lazy QR code (qrcode npm) and a
datetime-local "Läuft ab" picker. Both stay in sync with the active
URL — regenerating the link rebuilds the QR; clearing the expiry
re-publishes with no `expiresAt`.

Wired across all three unlisted collections:
- Calendar: LocalEvent.unlistedExpiresAt + setUnlistedExpiry +
  preserve-on-refresh + clear-on-flip; both Workbench DetailView and
  EventDetailModal pass expiresAt+onExpiryChange to SharedLinkControls.
- Library: same pattern in libraryEntriesStore + DetailView.
- Places: same pattern in placesStore + DetailView.

setVisibility clears any prior expiry so a flip-away-flip-back gets
a fresh "never expires" link. refreshUnlistedSnapshot and
regenerateUnlistedToken preserve the existing expiry so a content
edit or token rotation never silently extends a link's lifetime.

The qrcode dep ships as a regular `dependencies` entry on
@mana/shared-privacy so any consuming app picks it up via the
workspace.

Note: an unrelated svelte-check error in writing/components/DraftCard
("draft" not assignable to DragType) exists from a parallel session
and is not introduced by this commit.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:29:53 +02:00
Till JS
85fca7ccdc feat(comic): workbench app-registry — Comic als Card verfügbar
Comic war bisher nur als Standalone-Route /comic, /comic/new,
/comic/[id] erreichbar. Damit konnte man die App nicht via
AppPagePicker zur Workbench-Szene hinzufügen, das Comic-Autor-
Persona-Template (das `comic` in scene.openApps führt) hatte
keinen funktionierenden Card-Loader, und Cross-Module-Drops auf
das Comic-Modul gingen ins Leere.

Registrierung in app-registry/apps.ts:
- id: 'comic', name: 'Comic', color: '#f97316' (Orange-Rot
  Gradient analog mana-apps.ts)
- icon: FilmStrip — sequentielle Panels = Filmstreifen, passt zur
  Comic-Strip-Metapher und unterscheidet sich von Library (Stack)
  und Picture (Image)
- views.list: dynamic import von $lib/modules/comic/ListView.svelte
  (das Modul-Root mit Story-Grid + "+ Neue Story"-CTA)
- contextMenuActions: "Neue Story" → /comic/new (gleiche Pattern
  wie website's "Neue Website" — Hard-Navigation, weil der Create-
  Flow eine eigene Route ist)
- Detail-View nicht im Workbench-Slot wired — wie library/writing/
  wardrobe nutzt comic SvelteKit-Routes für Story-Detail, der
  Workbench-Card hostet nur die Liste

Help-Content-Eintrag (help-content.ts) für das ?-Icon im Card-
Header: Beschreibung der drei Generate-Modi, fünf Stil-Presets,
Modell-Auswahl (gpt-image-2 / Nano Banana), Cross-Modul-
Storyboard, MCP-Tools, Tipps zu Face-Ref + englischer Sprache +
Style-Fix-after-Create + 8-Panel-Konsistenz-Limit.

Comic-spezifische Files type-checken sauber. Die 240 Workspace-
weiten Errors sind pre-existing (laufender useLiveQueryWithDefault
→ useScopedLiveQuery Refactor).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:25:48 +02:00
Till JS
a80e8f57a0 feat(writing): auto-title suggestion in the briefing
The plan's open question on title-handling — "Topic = initialer Titel,
beim ersten Generate Auto-Title-Vorschlag anbieten" — answered with a
small -button next to the title input.

- prompt-builder: buildTitleSuggestionPrompt(input) returns a system+
  user pair that asks for a single 4–8-word title in the briefing's
  language. System prompt is strict: no quotes, no period, no "Titel:"
  prefix, no Markdown — so the result drops cleanly into the input
  field. cleanSuggestedTitle() strips wrapping quotes (straight + curly,
  single + double + German „"), a "Titel:" prefix, and a trailing
  period as a defense-in-depth pass.
- BriefingForm: -button next to the title input, disabled until the
  topic field has content (the suggestion needs context). On click it
  calls callWritingGeneration with the title prompt + temperature 0.6
  + maxTokens 60. In edit mode it pulls an excerpt of the current
  version (up to 800 chars) so the title hugs the actual prose, not
  just the briefing.
- The button shows a "…" while running and an error inline if the
  call fails — non-blocking, the user can still type their own title
  and save.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:16:45 +02:00
Till JS
636138b2d4 refactor(scope): replace _scopeCursor bridge with reactive useScopedLiveQuery hook
The previous fix wired Dexie's `_scopeCursor` infra-table as a side-
channel between Svelte $state (active-space + current-user) and Dexie
liveQuery: every scoped query touched the table on read so liveQuery
subscribed to it, every setActiveSpace bumped the table so liveQuery
re-ran. Worked, but smelled — hidden side-effect inside `scopedTable`,
scope state pretending to be a Dexie row, +1 roundtrip per query, and
`current-user.ts` had to dynamic-import Dexie just to pump the bridge.

Replacement: a Svelte 5 `$effect`-based hook that owns scope-tracking
explicitly. The dependency now lives in the reactive layer (which is
where it belongs), not as a side-effect in the data layer.

What changes:

- New `data/scope/use-scoped-live-query.svelte.ts`. The hook reads a
  module-level `scopeTick` `$state` counter inside its `$effect`.
  Both `onActiveSpaceChanged` (existing) and `onCurrentUserChanged`
  (new, added to `current-user.ts`) bump the tick on real changes.
  Effect re-fires → previous Dexie subscription unsubscribes → fresh
  one created with up-to-date `getInScopeSpaceIds()`. Same return
  shape as `useLiveQueryWithDefault` for drop-in migration.

- `current-user.ts` gains an `onCurrentUserChanged` event bus,
  symmetric to `active-space.svelte.ts#onActiveSpaceChanged`. Stays
  a plain `.ts` (no runes) so it remains a leaf and works in the
  test runner without the Svelte preprocessor — the rename to
  `.svelte.ts` was tried earlier and reverted because of test
  fallout (commit `01e6b9f04`).

- 53 module `queries.ts` files migrated from
  `useLiveQueryWithDefault` → `useScopedLiveQuery`. The choice of
  hook now documents at the call-site whether the query is
  scope-aware. Pure mechanical find-replace — no logic changes.

What goes away:

- `data/scope/cursor.ts` deleted.
- `touchScopeCursor()` calls removed from `scopedTable` /
  `scopedAnd` / `scopedGet`. Functions are pure data-layer again,
  no implicit reactive subscriptions.
- `bumpScopeCursor()` calls removed from `setActiveSpace` and both
  `loadActiveSpace` branches. Setter is pure state-update, no
  Dexie write side-effect.
- `current-user.ts` no longer dynamic-imports `scope/cursor` —
  `setCurrentUserId` is a clean three-line setter again.
- Dexie v46 drops the `_scopeCursor` table (`stores: { _scopeCursor:
  null }`). v45 stays declared so existing browsers' version chain
  remains contiguous; the v46 deletion runs once on next open. No
  user data lost — the table only ever held a transient bumpedAt
  row.

Existing 14 scope regression tests still pass. Type-check + theme-
token validators are clean.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:15:22 +02:00
Till JS
167d616cf7 feat(library,places): M8.4 — extend unlisted-share to two more modules
Calendar pilot proved the pattern in M8.3; this rolls it out to
Library entries and Places using the same backbone (resolvers
dispatcher, share-route SSR, SharedLinkControls UI).

Changes:
- lib/data/unlisted/resolvers:
    buildLibraryEntryBlob (whitelist: title, kind, creators, year,
      coverUrl, rating). Review, status, tags, progress, externalIds,
      reading-habit fields all stay private.
    buildPlaceBlob (whitelist: name, address, category). Lat/lng
      explicitly NOT inlined — 10m precision identifies homes /
      workplaces; the v1 share page renders no map. v1.1 may add an
      opt-in toggle.
    Dispatcher gains 'libraryEntries' + 'places' cases.
- modules/library: LibraryEntry gains unlistedToken; converter +
  ListView mock-stub forward it; entries store gets the same publish/
  revoke/refresh/regenerate quartet from M8.3:
    - setVisibility coordinates with mana-api server-side; failure
      aborts the local flip so Dexie + server stay aligned
    - deleteEntry revokes the active snapshot before tombstoning
    - updateEntry fire-and-forgets refreshUnlistedSnapshot so the
      shared link tracks edits to the whitelisted fields
    - regenerateUnlistedToken: revoke + republish, returns new token
- modules/library/views/DetailView: SharedLinkControls dropped into
  the existing dl as a labeled dt/dd row, only when visibility ===
  'unlisted' AND unlistedToken AND shareUrl.
- modules/library/SharedLibraryEntryView: standalone public render —
  big cover image, title, creators · year, optional rating-stars,
  OG/Twitter meta tags with cover as og:image (link-preview shows
  the cover on WhatsApp/Slack/iMessage).
- modules/places: same pattern. Place gains unlistedToken; converter
  + store get publish/revoke/refresh/regenerate; DetailView field-row
  for the SharedLinkControls.
- modules/places/SharedPlaceView: standalone public render — name,
  address, category badge, "Auf OpenStreetMap suchen"-Link (no map
  iframe in v1 because lat/lng aren't in the blob).
- routes/share/[token]/+page.svelte: dispatcher gains two more cases.

Verified:
- pnpm check (web): 7543 files, 0 errors, 0 warnings (svelte-check
  passes on all the new components, hooks, types)

Tests: vitest currently fails to load due to an unrelated parallel
edit on $lib/data/current-user.ts that uses $state in a non-.svelte.ts
file. The break predates this commit and isn't surfaced by
svelte-check; the visibility-system + unlisted code itself
type-checks clean. Will be fixed by whoever's currently iterating
on current-user (separate session).

Next: M8.5 — QR codes via the qrcode npm package, expiry-datepicker
wiring, regenerate confirm-dialog polish, and end-to-end
incognito-tab smoke test of all three modules' share links.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:13:36 +02:00
Till JS
bad935c258 feat(writing): print CSS + keyboard shortcuts
Two DetailView polish items.

Print / PDF (fixes M10's "Drucken / PDF" action):
- New <article class="print-target"> at the top of the route renders
  just the title + current version content. Hidden on screen, only
  visible under @media print so window.print() produces a clean
  manuscript instead of dumping the whole workbench chrome.
- :global(body > *) toggle suppresses the surrounding SvelteKit /
  workbench frame; the .shell + the per-card chrome are explicitly
  display:none in print. @page margin: 2cm gives a readable page
  with no further user setup.
- Body uses ui-serif so the printed prose looks like manuscript.

Keyboard shortcuts (DetailView document-level listener):
- ⌘G / Ctrl+G       → generate / re-generate (was: only the button)
- ⌘⇧S / Ctrl+Shift+S → save checkpoint
- ⌘Z / Ctrl+Z        → undo last refinement (only fires when refineUndo
                       is set; otherwise falls through to the textarea's
                       native undo as the user expects)

Buttons + the undo row carry the shortcut in their title attribute so
mouse-users discover them via tooltip.

i18n baseline +1 for DetailView (the new "(⌘Z)" tooltip suffix counts
as one additional German fragment per the validator's heuristic).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 12:10:32 +02:00
Till JS
01e6b9f044 fix(scope): undo accidental current-user.svelte.ts rename
A stray rename of current-user.ts → current-user.svelte.ts got swept
into the previous writing M5-expansion commit by lint-staged's stash/
restore step. The migration to a runes-bearing .svelte.ts variant
hadn't yet been completed (34 importers still use the old path), so the
result was a broken build until those importers were updated.

Renames the file back to current-user.ts and switches the one import
that had been pre-emptively rewritten (use-scoped-live-query.svelte.ts)
back to the bare path. The runes migration can land cleanly later.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 11:46:50 +02:00
Till JS
26e25b7694 feat(writing): M5 expansion — kontext, goal, me-image references
Three more reference kinds the resolver previously stubbed out are now
fully wired through the briefing form into the LLM prompt.

- reference-resolver: three new resolveX functions.
  * Kontext is a singleton per space (the picker uses a sentinel
    targetId; the resolver ignores it and reads via scopedForModule
    + first non-deleted row). Decrypts content and trims to budget.
  * Goal reads from companionGoals (plaintext today) and surfaces
    title + description + status + current/target so the model can
    tie the draft into the user's actual progress.
  * MeImage reads from the space-scoped meImages table; encrypts
    label + tags. Hands the model a textual descriptor (kind / label /
    tags) since the binary blob can't help prose generation.

- ReferencePicker: three new kind-tabs (🗂 Kontext, 🎯 Ziel, 🖼 Bild).
  Kontext renders as a single-click "Kontext-Dokument verknüpfen" entry
  if the space has one (with /kontext deep-link otherwise). Goals
  active-first, then archived/done. Me-images render with thumbnail +
  label + tags. Live-resolved chips via labelFor() for all three.

- i18n baseline bumped by one for ReferencePicker (the new
  "Kontext-Dokument verknüpfen" string is intentional, in line with
  the rest of the picker's existing German labels).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 11:43:12 +02:00
Till JS
fbbadc91f0 feat(calendar): M8.3 — calendar pilot for unlisted-share end-to-end
Wires Calendar through the M8.1+M8.2 backbone: flipping an event to
'unlisted' now publishes a server-side snapshot, the visible link in
the DetailView/EventDetailModal opens a real /share/[token] page, and
recipients can download an .ics file for their own calendar.

Changes:
- lib/data/unlisted/resolvers.ts (new):
    buildUnlistedBlob(collection, recordId) dispatcher.
    buildEventBlob: load LocalEvent + linked TimeBlock, decrypt
    client-side, return { title, location, startTime, endTime,
    isAllDay, timezone }. Description, reminders, tagIds, calendarId,
    color stay out of the blob — sensitive context the user didn't
    consent to share by flipping a single flag.
- modules/calendar/types: CalendarEvent gains `unlistedToken: string`
  (empty string when no active token). timeBlockToCalendarEvent
  forwards from LocalEvent. Draft-event scaffold initializes empty.
- modules/calendar/stores/events:
    setVisibility now coordinates with mana-api. Flip-to-unlisted:
      build blob -> publishUnlistedSnapshot -> store server-issued
      token in patch.unlistedToken -> commit local update. If the
      server call fails, no local change happens (no drift).
    Flip-from-unlisted: revoke server snapshot first, then clear
      local token + commit visibility change.
    deleteEvent: revoke active unlisted snapshot before tombstoning,
      so the share-link dies in lock-step with the local delete.
    updateEvent + updateSingleInstance fire-and-forget
      refreshUnlistedSnapshot(id) so the published blob tracks any
      whitelist-field edits. Failures log; the next successful
      refresh heals.
    New regenerateUnlistedToken(id): revoke + republish in one call,
      returns the fresh token. Powers the "Neu erzeugen" UI.
- routes/share/[token]/+layout.svelte: minimal anonymous chrome —
  no app nav, no auth, no Dexie. Light/dark via prefers-color-scheme.
  Footer carries "Geteilt via Mana" + signup CTA.
- routes/share/[token]/+page.server.ts: SSR loader. Fetches
  /api/v1/unlisted/public/:token, dispatches 404/410 cleanly,
  sets Cache-Control: private, max-age=60 + X-Robots-Tag: noindex.
- routes/share/[token]/+page.svelte: dispatcher; renders
  SharedEventView for collection='events', stub message otherwise.
- modules/calendar/SharedEventView.svelte: standalone public render —
  big date, location, "Zum eigenen Kalender hinzufügen" .ics link,
  optional expiry note. OG/Twitter meta tags for WhatsApp/Slack
  preview embedding. Uses $derived everywhere so prop updates
  propagate through reactive recompute.
- routes/share/[token]/ical/+server.ts: RFC 5545 builder. No npm
  library — small enough to inline. Escapes per spec, CRLF endings,
  DTSTART/DTEND swap between VALUE=DATE and UTC depending on isAllDay.
  Wrong-collection requests get 400.
- modules/calendar/views/DetailView (Workbench) + components/
  EventDetailModal (/calendar route): SharedLinkControls dropped in
  below the visibility row when event.visibility === 'unlisted'
  AND event.unlistedToken AND shareUrl computed. The URL is built
  client-side via buildShareUrl(window.location.origin, token) so it
  stays in sync with whichever host the editor is open on.

Verified:
- pnpm check (web): 7541 files, 0 errors, 0 warnings
- pnpm test calendar + website: 26/26
- typecheck of new resolver, store hooks, SSR loader, iCal builder

Manual test path:
1. Open /calendar event in Detail view, flip Sichtbarkeit -> "Per Link"
2. Server publishes snapshot, Dexie record gets the server token
3. SharedLinkControls appear with copy + regenerate + revoke buttons
4. Open the URL in incognito → SSR fetches snapshot, renders
   SharedEventView with date / location / .ics download
5. Edit the event title back in the main app → snapshot auto-refreshes
   (refreshUnlistedSnapshot fires after updateEvent succeeds)
6. Flip back to "Bereich" → snapshot revoked server-side; subsequent
   incognito reloads return 410 Gone

Next: M8.4 — same wiring for Library + Places. Uses the same
infra (resolvers dispatcher, share dispatcher) — just adds two new
buildXBlob functions, two SharedXView components, and the store
hooks.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 11:40:53 +02:00
Till JS
d8feef1149 fix(scope): bridge active-space / user changes to Dexie liveQuery
Modules mounted before the active-space bootstrap finished rendered
empty on first paint and stayed empty until some unrelated Dexie
write happened to wake the querier up. Reproducible in wardrobe:
open the module, the face-ref banner stays visible (even though
face-ref is set), the garment grid is empty; create a new garment
and suddenly the existing ones appear alongside it.

Root cause: Dexie's `liveQuery` only re-runs when a Dexie table it
read during evaluation is written to. `getInScopeSpaceIds()` reads
from plain Svelte `$state` (active-space.svelte.ts `active` +
current-user.ts `currentUserId`), which is invisible to Dexie's
change tracker. So:

1. User opens /wardrobe.
2. First querier runs. getActiveSpaceId() is still null (bootstrap
   hasn't resolved yet). getInScopeSpaceIds() returns
   [`_personal:guest`].
3. Existing rows are stamped `_personal:<userId>` or a real
   space-id — no match. Filter emits [].
4. Bootstrap resolves, setActiveSpace fires, `active = realSpace`.
   Svelte $state assignment: invisible to Dexie.
5. liveQuery never re-runs; user sees empty until a subsequent
   Dexie write forces a re-evaluation.

Fix: add `_scopeCursor`, a single-row infra table, as a Dexie
proxy for the scope-state signal.

- `data/scope/cursor.ts`: new module with two exports.
  `bumpScopeCursor()` writes `{id:'active', bumpedAt}` to the
  cursor. `touchScopeCursor()` reads from it, fire-and-forget —
  the Dexie read registers the liveQuery subscription during
  querier execution.
- Dexie v45 registers `_scopeCursor: 'id'`. NOT in SYNC_APP_MAP:
  it's a client-side liveness signal, not user data, so the
  creating-hook loop ignores it and no pending-change rows are
  generated.
- `scopedTable` / `scopedGet` / `scopedAnd` in `scoped-db.ts` call
  `touchScopeCursor()` on every invocation. Since those run inside
  each querier's body, liveQuery picks up the subscription for free.
- `setActiveSpace` (direct + both loadActiveSpace branches) calls
  `bumpScopeCursor()` after the `$state` update.
- `setCurrentUserId` in current-user.ts calls `bumpScopeCursor()`
  via dynamic import so the module stays leaf-level (no eager
  Dexie dep in test envs that mock fake-indexeddb differently).

Net effect: every scope change triggers one Dexie write to an
infra table, every liveQuery that went through `scopedForModule`
sees the write and re-evaluates with the fresh
`getInScopeSpaceIds()`. The first-mount race that left wardrobe
stuck on an empty list is gone.

Existing 14 scope regression tests still pass.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-25 11:32:32 +02:00
Till JS
3f438cf882 refactor(i18n): replace hand-maintained locale list with import.meta.glob
Bisher war für jeden Namespace ein eigener `import(\`./locales/X/\${lang}.json\`)`
Eintrag in index.ts + ein spiegelndes Destructuring — 35 Namespaces,
zwei hand-synchronisierte Listen, jeder neue Namespace erforderte Edits.
Leicht zu vergessen, lautlos fehlerhaft.

Vite's import.meta.glob übernimmt das jetzt: alle ./locales/*/*.json
werden automatisch für den gewünschten Lang registriert, Namespace aus
dem Pfad abgeleitet. Neuer Namespace = JSON-Folder droppen, Ende.

- index.ts: 118 Zeilen → 15 Zeilen.
- svelte-check 0/0.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:21:43 +02:00
Till JS
e794e0ca57 feat(i18n): locale-aware formatters, migrate hardcoded de-DE call-sites
Bisher pinnten 185+ call-sites die UI-Zahlen/Datumsausgabe hart auf
Deutsch ("de" / "de-DE" an toLocaleDate{,Time}String, Intl.*Format,
date-fns locale), unabhängig von der aktiven Sprache. EN-User sahen
dadurch deutsche Datums-/Zahlenformate mitten im englischen UI.

- $lib/i18n/format.ts (neu): formatDate / formatTime / formatDateTime
  / formatNumber / formatCurrency / getDateFnsLocale. Alle lesen die
  aktive Locale aus svelte-i18n's locale-store und mappen de→de-DE
  etc. für Intl.
- Codemod: 119 Direktaufrufe in 79 Files migriert (.toLocaleDateString
  / .toLocaleTimeString / new Intl.{Number,DateTime}Format). Erkennt
  new Date() / Number()-Receiver zum Disambiguieren von
  .toLocaleString('de-DE').
- date-fns: 19 Files auf getDateFnsLocale() umgestellt; hardcoded
  `import { de } from 'date-fns/locale'` entfernt.
- Skipped (Collision): 14 Files hatten lokale format*-Wrapper; diese
  bleiben vorerst als gesonderte Folge-Refactorings stehen. Ca. 58
  deep-gekapselte Aufrufe im .toLocaleString/.toLocaleDateString-Idiom
  sind über diese Wrapper noch zu migrieren.
- svelte-check: 0 Errors / 0 Warnings.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:20:11 +02:00
Till JS
364522db87 feat(comic): image-model picker — OpenAI + Nano Banana wählbar
Comic nutzte bisher 'openai/gpt-image-2' hartcodiert auf drei Ebenen
(generate-panel.ts, comic.generatePanel MCP-Tool, generate_comic_panel
AI-Tool). Wardrobe hat seit dem Nano-Banana-Commit einen
TryOnModelPicker mit drei Optionen — Comic spiegelt das jetzt 1:1.

Wählbar in allen drei Editoren (PanelEditor, BatchPanelEditor,
StoryboardSuggester):
- openai/gpt-image-2 (Default) — OpenAI GPT-image Standard
- google/gemini-3-pro-image-preview — Nano Banana Pro, hohe
  Konsistenz, teurer
- google/gemini-3.1-flash-image-preview — Nano Banana 2, neuestes,
  schnell, günstig

Implementierung:
- api/generate-panel.ts: PanelModel Union + DEFAULT_PANEL_MODEL +
  model? Param auf RunPanelGenerateParams + im HTTP-Body
  weitergereicht (vorher hart 'openai/gpt-image-2').
- components/PanelModelPicker.svelte: neue Komponente, Stil/Markup
  identisch zu TryOnModelPicker für Muskel-Memory über beide Flows.
- components/PanelEditor.svelte: `let model = $state(DEFAULT_PANEL_MODEL)`
  + Picker oberhalb der Qualität-/Format-Leiste + model im
  runPanelGenerate-Call.
- components/BatchPanelEditor.svelte: gleiche Änderung — ein Model
  pro Batch (nicht pro Row) damit der Batch konsistent rendert.
- components/StoryboardSuggester.svelte: gleiches Pattern; der
  Picker landet zwischen "Panel manuell"-Button und dem
  Qualität/Format-Block.
- packages/mana-tool-registry/src/modules/comic.ts: generatePanel
  Input-Schema bekommt model mit zod.enum() + default; im Body
  wird input.model durchgereicht.
- packages/shared-ai/src/tools/schemas.ts: generate_comic_panel
  bekommt Parameter 'model' optional mit gleicher Enum-Liste.
- apps/mana/apps/web/src/lib/modules/comic/tools.ts: isValidModel
  Guard + Parameter-Validierung; model an runPanelGenerate.

Keine Story-Level-Persistierung — model bleibt lokaler State pro
Editor-Mount. Eine model-Spalte auf comicStories würde Migration
brauchen und die Wahl ist eh ad-hoc pro Panel/Batch.

Plan-Doc (§2.1) dokumentiert die Entscheidung + die drei Optionen.

107 shared-ai tests weiter grün. check + validate:all clean.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:19:40 +02:00
Till JS
c1d643ffb5 feat(writing): register Writing as a Workbench card
Writing is now pickable from the Workbench AppPagePicker and can run
side-by-side with the other module cards. Previously only reachable via
direct navigation to /writing.

- app-registry: registerApp({ id: 'writing', icon: NotePencil, color:
  '#0ea5e9', list: ListView }) with a "Neuer Draft" context-menu action
  that dispatches mana:quick-action (same convention as library/notes/
  quiz — the event channel is the standard cross-card hook).
- writing ListView: onMount listener for the mana:quick-action event
  flips showCreate=true when { app:'writing', action:'new' } fires, so
  the kebab menu → "Neuer Draft" opens the inline BriefingForm instead
  of being a no-op.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:15:25 +02:00
Till JS
92bee0d71a feat(unlisted): M8.1 — backend foundation for shareable-link snapshots
First milestone of the unlisted-share rollout plan (docs/plans/
unlisted-sharing.md). Adds the server-side infrastructure that backs
`visibility='unlisted'` — previously the flag was stamped locally but
led nowhere. After this commit, a token points at an actual snapshot
the SSR share-page will render (M8.3+).

Scope: backend only. No client-side publish/revoke calls yet, no
share-route, no UI. That lands in M8.2/M8.3. Anyone hitting the
endpoints manually with curl can exercise the full publish-fetch-
revoke cycle.

Changes:
- New pgSchema `unlisted` with table `snapshots`:
    token (pk, 32-char base64url)
    user_id, space_id, collection, record_id, blob (jsonb)
    created_at, updated_at, expires_at (nullable), revoked_at
  Partial unique index on (user_id, collection, record_id) WHERE
  revoked_at IS NULL so one record has at most one active token.
  Partial btree on expires_at for the cron-cleanup path.
- Hand-authored SQL migration `apps/api/drizzle/unlisted/0000_init.sql`
  (manual-apply per the repo's feedback_api_hand_authored_migrations
  memory). Already applied to the local mana_platform.
- Drizzle schema `apps/api/src/modules/unlisted/schema.ts`. All id
  fields are `text` not uuid — Better-Auth nanoids aren't UUIDs, same
  trap we hit with the website module's publish bug.
- mana-api module `apps/api/src/modules/unlisted/`:
    POST   /api/v1/unlisted/:collection/:recordId (auth)
      Body: { spaceId, blob, expiresAt? }. Re-publish reuses the
      existing active token (by (user,collection,record) lookup); a
      revoke-then-republish mints a fresh token row. Response includes
      a fully-qualified share URL built from Origin/Referer/env.
    DELETE /api/v1/unlisted/:collection/:recordId (auth)
      Soft-revoke. Idempotent — already-revoked returns
      { revoked: 0 } cleanly so client stores can call it
      unconditionally on setVisibility-away.
    GET    /api/v1/unlisted/public/:token (public)
      Rate-limited 20/min/token + 60/min/ip so token enumeration is
      impractical. 404 for unknown, 410 Gone for revoked or expired.
      Cache-Control: private, max-age=60 + X-Robots-Tag: noindex for
      SEO isolation. Returns { token, collection, blob, createdAt,
      updatedAt, expiresAt }.
- ALLOWED_COLLECTIONS hardcoded allowlist in POST handler
  (events, libraryEntries, places — the M8.3+M8.4 scope). Unknown
  collection -> 400 COLLECTION_NOT_ALLOWED. Keeps the schema honest
  about what the server accepts.
- drizzle.config extended to include the new schema in managed
  migrations.
- index.ts wires unlistedPublicRoutes pre-auth (before
  authMiddleware) and unlistedRoutes post-auth.

Verified:
- Migration applied to mana_platform — `unlisted.snapshots` exists
  with both partial indexes.
- pnpm run type-check (api): clean
- pnpm run validate:all: theme-tokens, theme-parity, crypto-registry,
  encrypted-tools all green
- URL build uses Origin/Referer before the env fallback so dev
  (http://localhost:5173) and prod (https://mana.how) both work
  without env churn.

Next: M8.2 — shared-privacy client helper + SharedLinkControls
component.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 17:12:13 +02:00
Till JS
6f37e00bf4 feat(comic): AI_TOOL_CATALOG bridge — webapp-runner kann jetzt Comics
Macht den Comic-Autor-Template (M6) auch im Web-App-Mission-Runner
nutzbar. Bisher war der Template nur über persona-runner/Claude
Desktop sinnvoll, weil die comic.*-Tools nur im mana-tool-registry
(MCP) lagen. Jetzt kennt die AI Workbench drei neue Tools und der
Template-Policy-Map trägt beide Naming-Konventionen.

AI_TOOL_CATALOG-Einträge (packages/shared-ai/src/tools/schemas.ts):
- list_comic_stories (auto) — filter style?/favoriteOnly?/limit?
- create_comic_story (propose) — title + style + optional
  description/storyContext/tags. Character-Refs werden vom Executor
  automatisch aus meImages primary face-ref + body-ref gezogen,
  also muss der Planner keine mediaIds kennen.
- generate_comic_panel (propose) — storyId + panelPrompt + optional
  caption/dialogue + quality. Kostet Credits.

Executors (apps/mana/apps/web/src/lib/modules/comic/tools.ts):
- list: scopedForModule pull + decrypt + filter + sort newest.
- create: resolveCharacterMediaIds() scannt meImagesTable für das
  aktive Space, nimmt face-ref+body-ref. Fehler wenn kein Face
  hinterlegt ("Lade eines in /profile/me-images hoch"). Delegiert
  an comicStoriesStore.createStory — gleiche encryption/event-
  pipeline wie StoryForm.
- generate: lädt Story decrypted, delegiert an runPanelGenerate
  (identischer Pfad wie PanelEditor in der UI), liefert
  panelIndex + imageUrl zurück.

Registrierung in data/tools/init.ts (registerTools(comicTools)).

Template-Policy (comic-author.ts) jetzt bi-lingual: snake_case
(AI_TOOL_CATALOG) UND dot-case (MCP) nebeneinander in tools-Map.
So gilt die Intent-Policy konsistent egal welche Runner-Oberfläche
das Tool nennt — auto für list_comic_stories / comic.listStories,
propose für create_comic_story / comic.createStory /
generate_comic_panel / comic.generatePanel / comic.reorderPanels.

apps/mana/CLAUDE.md Tool-Coverage-Tabelle bekommt eine Comic-Zeile.

Tool-Count jetzt 75→78, Module 22→23. 107 shared-ai tests
weiter grün. check + validate:all clean.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:49:24 +02:00
Till JS
6545498dc2 feat(writing): agent.defaultWritingStyleId — M8 persona-linkage follow-up
Agents can now pin a default writing style. When an AI-actor runs
`create_draft` without an explicit styleId, the tool resolves to the
agent's `defaultWritingStyleId` so e.g. a "Marketing-Agent" always
drafts in the Corporate-Tone style and a "Memoir-Agent" in Memoir.

- @mana/shared-ai: optional `defaultWritingStyleId?: string` added to
  the Agent interface (plaintext FK, format `preset:<id>` or a custom
  WritingStyle uuid). No migration — existing rows stay undefined and
  the fallback path no-ops for them.
- ai-agents store: field threaded through CreateAgentInput + AgentPatch
  + the create function's copy-list. `updateAgent` already deep-clones
  the patch so nothing else to change there.
- ai-agents ListView: new "Writing" section in the agent detail panel
  with a StylePicker (reuses the writing module's component — Vorlagen
  + Meine Stile optgroups). Empty = kein Default.
- writing/tools.ts: `resolveAgentDefaultStyle()` reads the current
  actor, guards `isAiActor`, loads the agent row, and returns its
  defaultWritingStyleId. Wired into `create_draft` as a fallback when
  `params.styleId` is missing. User-invoked calls skip the lookup — a
  human omitting styleId means "ad-hoc, no style", not "my default".
  `generate_draft_content` needs no change because the draft's styleId
  is already set at create time.

107 shared-ai tests still pass.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:36:20 +02:00
Till JS
e7398b2dee polish(wardrobe): make Try-On CTA the dominant element + shrink secondary actions
Feedback from the latest screenshot: on the dark workbench theme the
earlier "lift + shadow" polish was too subtle — the Try-On button
read as equal weight to "Heute getragen" / "Archivieren" / "Löschen",
so it didn't telegraph "this is THE action on the page". The
primary-tinted shadow I used (shadow-primary/20) didn't have enough
contrast against the dark card.

Stronger visual hierarchy:

- CTA bumps to rounded-xl + px-6 py-4 + text-lg + font-semibold (was
  rounded-lg + px-5 py-3.5 + text-base + font-medium).
- At rest: neutral shadow-lg + primary/40 ring-2 so the button has a
  visible glow on both dark and light themes (neutral shadow works
  regardless of bg contrast; the primary ring carries the accent).
- Hover: lift -translate-y-0.5 + shadow-xl + ring-4 at primary/50.
- Active-press: back to baseline (translate-y-0, shadow-md, ring-2)
  for tactile feedback.
- Sparkle icon 18 → 20; gap tightens to 2.5.
- Applied identically to both GarmentTryOnButton + TryOnButton so
  solo and outfit surfaces share the same CTA weight.

Secondary-action row collapses from three equal full-width buttons
into one labelled button + two icon-only buttons:

- "Heute getragen" keeps the full label and takes the remaining row
  width — it's the frequent positive action.
- "Archivieren" + "Löschen" shrink to 36×36 icon buttons with
  tooltips (aria-label + title) carrying the full label. Archive
  uses the primary-tinted hover; Löschen keeps its destructive-red
  hover.

Net effect: the Try-On CTA is now clearly the loudest element on the
page, Heute-getragen sits cleanly below as a secondary action, and
Archive/Löschen recede to minimal icon-only controls — no more
visual competition with the primary CTA.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:35:18 +02:00
Till JS
0ee3b145f0 polish(wardrobe): unify hover vocabulary + lift the Try-On CTA
The garment detail page used three different hover dialects — the
model picker reacted with a primary-tinted bg + border (feels like a
button), the secondary action buttons had a plain muted-grey hover,
the edit pencil was invisible until hover, and the hero photo was
entirely static. Result: the model picker was the only place that
telegraphed "click me". Everything else felt flat.

Align on one vocabulary across the page: primary-tinted border +
primary/5 bg on hover for anything interactive.

- Hero photo is now a `<button>` that opens the existing ImageLightbox
  with the garment's full-res mana-media URL (synthesised as a minimal
  picture.Image — prompt=name, no model/dims/date noise). Hover adds
  the primary-tinted border + a subtle shadow-md + a 1% scale on the
  `<img>` for depth.
- Edit pencil becomes a labelled button "Bearbeiten" with the same
  primary hover. No more hover-to-discover — editing reads as a
  first-class action.
- "Heute getragen", "Archivieren" drop the plain muted hover for the
  primary-tinted one. "Löschen" keeps its destructive-red tint but
  adds border-error/50 on hover so it feels as interactive as the
  others.

Try-On CTA now reads as the most important action:
- rounded-lg + px-5 py-3.5 + text-base + font-semibold (was
  rounded-md + px-4 py-2 + text-sm + font-medium).
- shadow-md shadow-primary/20 at rest → shadow-lg shadow-primary/30
  on hover, combined with -translate-y-0.5 for a subtle lift.
- active:translate-y-0 + shadow-sm makes the press feel tactile.
- Sparkle icon bumped 16 → 18, spinner likewise.

Applied to both GarmentTryOnButton (solo) and TryOnButton (outfit)
so the two surfaces share CTA weight.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:31:11 +02:00
Till JS
9c503b7982 i18n: strip dead legacy keys + fill IT/FR/ES calendar translations
Vor-Audit: 4 Namespaces mit Drift (calendar, dashboard, memoro, times).
Nach diesem Commit: alle 35 Namespaces × 5 Locales in voller Parität.

Tote Keys entfernt (0 Call-Sites laut grep über src/):
- memoro/{it,fr,es}.json: app_slider.* (16 Keys) — wurde längst in
  den app_slider/-Namespace migriert.
- calendar/*.json: sync.*, sharing.* (aus allen 5), auth.*, event.save,
  event.delete, event.cancel (aus en/it/fr/es — DE war schon sauber),
  calendar.common.* (aus it/fr/es) — alles duplizierte common-Strings.
- times/{de,en}.json: auth.* (6 Keys) — nur dort als Dupe geblieben.

Echte Lücken gefüllt:
- calendar/{it,fr,es}.json: 15 Keys (views.weekNumber, views.moreEvents,
  calendar.draftEvent/hideSidebar/showSidebar, event.changeStartTime/
  EndTime, errors.{load,create,update,delete}Events, success.event*,
  a11y.createEventOn/slotTime) — waren in DE/EN vorhanden, in IT/FR/ES
  via fallbackLocale auf Deutsch zurückgefallen.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:28:10 +02:00
Till JS
05b2209232 polish(wardrobe): garment-detail cosmetic pass + slug-cleanup on upload
Four small UI tweaks that came out of reviewing the garment-detail
screenshot against the workbench chrome:

1. Duplicate "Kleiderschrank" label — the ModuleShell header above
   DetailGarmentView already renders a back-arrow and the app title.
   The inner `<nav>` with a second arrow + text was rendering it all
   a second time. Drop the inner breadcrumb; ArrowLeft import along
   with it.

2. Raw SKU-slug as default garment name — the old
   `stripExt(file.name)` produced labels like
   `17390-gestreiftes-herren-t-shirt-aus-baumwolle-17390-2-w`. New
   `prettifyUploadName` helper:
   - drops the extension
   - replaces `-`/`_` with spaces
   - strips pure-digit tokens of length ≥ 4 (SKU shape) but keeps
     short alphanumerics like `4xl` / `w38`
   - title-cases each remaining word, rebuilding hyphens
     (`t-shirt` → `T-Shirt`, `v-neck` → `V-Neck`)
   - clamps to 80 chars on a word boundary
   GridView's ingestFiles now passes the prettified name into the
   createGarment write. User still edits on the detail page for
   anything that needs nuance.

3. Two-line CTA with Credits subtitle. The button used to read
   `Anprobieren · 10 Credits` on one line; on a narrow workbench
   card the mittelpunkt between label and cost was visually thin
   and read like a strikethrough. Split into a main label + small
   opacity-75 subtitle so the credit figure is clearly secondary
   info, not a decorated part of the CTA text. Applied to both
   GarmentTryOnButton and TryOnButton.

4. Redundant microcopy under section headers — "Einzelstück auf dir
   gerendert" under ANPROBEN and "Komposition öffnen" under IN
   OUTFITS repeated what the section title and the clickable cards
   already signalled. Remove both.

No behaviour changes, no schema, no API.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:24:07 +02:00
Till JS
87b567eec9 i18n: fix IT/FR/ES parity gaps in dashboard + memoro
- dashboard: +5 Einträge pro Sprache für die beiden neuen Widgets
  activity_feed + articles_unread.
- memoro: +1 Eintrag pro Sprache für memo.load_more.

Damit sind dashboard (111) und memoro auf gleichem Stand wie DE/EN.
Verbleibende Drift (app_slider-Legacy-Keys in memoro IT/FR/ES,
common/auth-Legacy in calendar/times) ist strukturell und bleibt
einem Folge-Cleanup vorbehalten.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:19:59 +02:00
Till JS
d49ad239d9 feat(writing): M8 — AI tools exposed through the shared catalog
Writing is now programmatically accessible from the foreground mission
runner, personas, and Claude Desktop / MCP. Eight tools land:

Auto (read-only):
- list_drafts         — filtered by kind/status + word-count summary
- get_draft           — briefing + current version body, ready for reading
- list_writing_styles — 9 presets + user customs, ids usable in create_draft

Propose (human approval per agent policy):
- create_draft            — briefing only, no generation yet
- generate_draft_content  — wraps generationsStore.startDraftGeneration;
                            writes a new LocalDraftVersion + pointer flip
- refine_draft_selection  — wraps refineSelection + applyRefinement in
                            one call; operations: shorten/expand/tone/
                            rewrite/translate with op-specific params
- set_draft_status        — draft/refining/complete/published
- save_draft_as_article   — hand-off to articlesStore.saveFromExtracted
                            with internal://writing/<id> as originalUrl,
                            records publishedTo + emits WritingDraftPublished

Schemas live in @mana/shared-ai/src/tools/schemas.ts (the SSOT that the
web-app policy layer + mana-ai planner derive from). Executors live in
modules/writing/tools.ts and delegate to the existing stores so the
encryption + event pipeline runs once regardless of who called the tool.
Registration added to data/tools/init.ts.

107 shared-ai tests still pass. CLAUDE.md tool-coverage table bumped:
67→75 tools, 21→22 modules.

Not in M8 (deferred): agent.defaultWritingStyleId linkage (needs a
Persona schema change + runner wiring), mana-tool-registry Zod specs
(add when a non-web MCP client needs richer validation).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:19:30 +02:00
Till JS
0dadc17771 feat(writing): M10 — publish hooks (export + save-as-article)
Writing can now leave the module. Four outbound paths in this milestone:
Markdown to clipboard, plain text to clipboard, .md file download, and
a browser-native print / PDF. Plus a first cross-module hand-off: save
as a read-later article.

- utils/export.ts: pure helpers — draftToMarkdown (title as H1 + body),
  draftToPlainText, downloadFile via synthetic anchor + blob URL,
  fileStem (NFKD-normalise + slug-ify the title), and a
  copyTextToClipboard wrapper that falls back to document.execCommand
  for http contexts that can't touch navigator.clipboard.
- draftsStore.recordPublish(draftId, module, targetId) — idempotent
  per (module, targetId), appends to draft.publishedTo with a
  publishedAt timestamp, emits WritingDraftPublished so the Workbench
  timeline picks up the hand-off.
- ExportMenu.svelte: dropdown next to Generate / Checkpoint with five
  items. "Als Artikel speichern" calls articlesStore.saveFromExtracted
  with originalUrl='internal://writing/<draftId>' as both a dedupe-safe
  identifier and a back-reference to the source draft, then records
  the publish and navigates to the new article.
- DetailView surfaces draft.publishedTo as green chips under the meta
  row, with a click-through to the articles reader. Only 'articles' has
  a landing page today; website / presi / mail / social-relay chips
  render the label without a link (they arrive with their own
  milestones).

Not in M10: website-block publishing (M-website), presi import
(M-presi), mail draft creation. Those each need per-target block /
slide / message shapes that exceed this commit's scope.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:07:39 +02:00
Till JS
6432ef7e6b feat(comic): M4 — AI-Storyboard aus Cross-Modul-Text
User wählt einen bestehenden Text (Tagebuch-Eintrag, Notiz oder
Bibliotheks-Review), das Modell schlägt eine geordnete
Panel-Sequenz vor (prompt + optional caption + dialogue pro Panel),
der User prüft/editiert und feuert Batch-Gen mit sourceInput-
Tagging — damit wird `useStoriesByInput` später cross-referenzieren
können ("Welche Comics sind aus diesem Journal-Eintrag entstanden?").

Backend:
- POST /api/v1/comic/storyboard (Hono route) nimmt style +
  sourceText + panelCount (+ optional storyContext / sourceModule)
  und ruft llmJson() mit einem response_format=json_object-Prompt
  an mana-llm. System-Prompt instruiert das Modell auf eine exakte
  {panels: [{prompt, caption?, dialogue?}]}-Shape, Rules wie
  "keine Style-Instruktionen" (kommen aus dem Story-Prefix
  downstream) und "kein Panel-Nummerieren".
- Defense-in-depth Coerce auf der Response: Panel ohne prompt
  wird gefiltert, Strings werden gecappt (caption/dialogue 200,
  prompt 800), Zahl der Panels auf panelCount geclampt.
- Model via COMIC_STORYBOARD_MODEL env var überschreibbar;
  Default ollama/gemma3:4b wie writing (lokal + billig).
- Beide Erfolgs- und Fehler-Pfade mit logger.info /
  logger.error + userId + sourceModule für Observability.
- Route registriert in apps/api/src/index.ts als /api/v1/comic.

Client:
- api/storyboard.ts: suggestPanels({style, sourceText, panelCount,
  storyContext?, sourceModule?}) — thin fetch-Wrapper + Error-Messaging
  für 402 / 502 / no-panels-Responses.
- ReferenceInputPicker: Tabs über Journal / Notizen / Bibliothek
  (die drei inhalts-dichtesten Quellen), pro Tab Live-Query +
  Suche + Entry-Liste. Click emittiert {module, entryId, label,
  sourceText} — label ist der Display-Name für die
  "Gequellt aus…"-Chip, sourceText ist bereits decrypted (Queries
  liefern plaintext zurück). Bibliotheks-Einträge ohne Review
  sind disabled (kein Text = nichts zu rendern).
- StoryboardSuggester: 4-Schritt-Flow (pick-source →
  generating-plan → review-plan → rendering). Schritt 3 ist der
  eigentliche Editor: jede Claude-Zeile ist editierbar (Prompt,
  Caption, Dialog) mit Trash-Button; Quality + Format-Toggle
  teilen sich M3-Batch-Style. "Generieren" ruft parallel
  runPanelGenerate() via Promise.allSettled mit
  sourceInput={module, entryId} im panelMeta, alle Panels gehen
  durch den identischen M2-HTTP-Pfad.
- DetailView bekommt einen dritten Editor-Modus "ai" neben
  "single" und "batch" — eine Sparkle-Button-CTA öffnet den
  Suggester.

Kein Writing-Draft / Calendar-Event-Input in dieser Runde —
Drafts brauchen Version-Chain-Resolve, Events sind meist zu dünn
an Prosa. Follow-up wenn gewünscht (rein additiv: Tab + Hook).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:06:03 +02:00
Till JS
8a882a3760 feat(wardrobe,picture): Google Nano Banana as a Try-On option
Add Google's Gemini image edit family (Nano Banana) as a user-
selectable model for Wardrobe Try-On next to the existing OpenAI
path. Three concrete choices now expose themselves in the Solo and
Outfit Try-On buttons:

  - openai/gpt-image-2          (default, falls back to gpt-image-1
                                 server-side when the org isn't
                                 verified)
  - google/gemini-3-pro-image-preview   (Nano Banana Pro — premium
                                 identity / character consistency)
  - google/gemini-3.1-flash-image-preview (Nano Banana 2 — newest,
                                 fast, cheapest)

All three accept multi-image refs (face + body + garment) through
the same /api/v1/picture/generate-with-reference endpoint; the only
differences are the provider-specific request/response shape and
the model-id routing.

Server (apps/api/src/modules/picture/routes.ts):
- Guard now accepts `openai/*` and `google/*` prefixes and rejects
  everything else as "not supported for edits". Each provider's key
  is validated separately so missing GEMINI_API_KEY doesn't break
  OpenAI calls and vice versa.
- New `callGeminiEdits(modelName)` helper mirrors the shape of
  callOpenAiEdits: encodes the normalized PNG refs as base64
  inline_data parts, POSTs to
  generativelanguage.googleapis.com/v1beta/models/{model}:generateContent
  with responseModalities=["TEXT","IMAGE"] and imageConfig
  (aspectRatio + imageSize), pulls the generated image out of
  candidates[].content.parts[].inlineData.
- Our internal size strings map cleanly: 1024x1024 → 1:1 / 1K,
  1024x1536 → 2:3 / 1K, 1536x1024 → 3:2 / 1K. Gemini 1K is enough
  for the thumbnail sizes Wardrobe renders; going higher bloats
  payload without visible gain.
- creditsFor() gains a google/ branch proportional to upstream
  pricing (pro ≈ 18, 3.1-flash ≈ 6, 2.5-flash ≈ 5).
- Response `model` reports `${provider}/${modelUsed}` so the picture
  row's model metadata is accurate across providers.

Client (apps/mana/apps/web/src/lib/modules/wardrobe):
- api/try-on.ts: export `TryOnModel` union + `DEFAULT_TRY_ON_MODEL`.
  RunGarmentTryOnParams / RunOutfitTryOnParams gain an optional
  `model` field, threaded through `callGenerateWithReference`.
- components/TryOnModelPicker.svelte: new segmented control, three
  options with label + one-line hint. Grid-auto-fits so it reflows
  on the narrow workbench card.
- components/GarmentTryOnButton.svelte + TryOnButton.svelte: both
  mount the picker above the Sparkle CTA. `estimatedCredits` on the
  button label updates live when the user switches model so the
  cost signal matches what the server will actually charge.

Env (scripts/generate-env.mjs): GEMINI_API_KEY and GOOGLE_API_KEY
now propagate from the root `.env.development` into `apps/api/.env`
so mana-api can pick them up at boot. The route reads GEMINI_API_KEY
with GOOGLE_API_KEY as fallback, matching how mana-llm ships today.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:04:21 +02:00
Till JS
90915b7879 fix(visibility): promote picker to labeled rows across remaining modules
Pattern established in fix(calendar, 8c43c119e) — the picker reads as
a first-class property when it sits in its own row with a 'Sichtbarkeit'
label instead of being tucked behind a compact icon. Applying the same
treatment everywhere the picker was still using the compact variant.

Changes:
- library/views/DetailView: move the picker out of the meta-top-row
  (kind-pill + picker cluster) and into the existing <dl class="details">
  block as a first dt/dd pair. Keeps the kind-pill standalone and gives
  visibility equal weight to the other structured details. Removes the
  now-orphaned .meta-top-row CSS rule
- wardrobe/views/DetailOutfitView: remove the compact picker from the
  header action cluster (it was competing with favourite/edit buttons);
  replace with a 'Sichtbarkeit' label + full picker in its own flex row
  directly below the header, above description
- calendar/components/EventDetailModal: remove the compact picker from
  the modal-actions row (copy/edit/delete chevron area). Add a new
  detail-row at the top of event-details with a 'Sichtbarkeit' label
  (new .detail-label CSS rule, mirrors .detail-icon layout for visual
  consistency with Time/Location/etc rows)

Picture board detail stays as-is: the picker already renders with its
label visible (non-compact) in the header flex, and the Board-Detail
page has no prop-row-style content list to slot it into — inline in
the header is the right spot there.

Verified:
- pnpm check (web): 7520 files, 0 errors, 0 warnings

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:03:48 +02:00
Till JS
f2275f752d feat(writing): M11 — adopt the unified visibility system
The backing for visibility was already in place from M2 (draft.visibility
stamped on create via defaultVisibilityFor, draftsStore.setVisibility
mints/clears unlistedToken and emits VisibilityChanged), so M11 is just
the UI step that puts it in front of the user.

- <VisibilityPicker> from @mana/shared-privacy sits in the meta-row of
  DetailView, mirroring the library pattern. onChange calls
  draftsStore.setVisibility — no new store method needed.
- Draft type + toDraft converter now surface `unlistedToken` so the UI
  can render a share row when visibility === 'unlisted'. Token is
  displayed verbatim + "Kopieren"-button because the public read-URL
  for drafts ships with M10 (Publish-Hooks); a tooltip makes that
  explicit so the user doesn't expect a working link yet.

With this, Writing is now consistent with the Library / Picture /
Calendar / Todo / Goals / Places / Recipes / Wardrobe pilot group.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:01:05 +02:00
Till JS
ed8ec98572 feat(workbench): app picker searches English aliases + IT/FR/ES i18n
Die Suche im AppPagePicker matcht jetzt gegen DE + EN + Fallback-Name +
id-Slug, damit englische Begriffe immer greifen — auch wenn die UI in
Deutsch läuft ("cal" → Kalender, "weather" → Wetter).

- AppPagePicker: statische Imports von apps/de.json + apps/en.json,
  neue searchHaystack-Funktion, Filter über Haystack statt nur
  displayName.
- apps/{it,fr,es}.json: +42 Einträge pro Sprache, alle 78 registrierten
  Module sind jetzt nativ übersetzt. Vorher fiel IT/FR/ES für neuere
  Module via fallbackLocale auf Deutsch zurück (z.B. "Wetter" im
  italienischen Menü).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:00:27 +02:00
Till JS
63f51799b0 feat(comic): M3 — Batch-Panel-Generierung (2–4 parallel)
Dem DetailView einen zweiten Editor-Modus spendiert: neben dem
Einzel-Button ("+ Panel") gibt es jetzt einen "+ Batch"-Button, der
`BatchPanelEditor` öffnet. Der Batch-Editor zeigt 2-4 Prompt-Cards
(prompt + caption + dialog pro Zeile, dynamisch per + / trash),
feuert alle Zeilen parallel über `Promise.allSettled` an
`runPanelGenerate`, und rendert pro Zeile Live-Status-Chips
(pending / ok / error) plus Retry-Button bei Fehlern.

- Parallel statt seriell, damit OpenAI-Latenz nicht N× addiert wird.
  `Promise.allSettled` isoliert jeden Call, ein 402-Credits-Fehler
  auf Zeile 2 bricht Zeilen 3-4 nicht ab.
- Nach erfolgreichem Submit werden die erfolgreichen Zeilen
  verworfen; fehlgeschlagene bleiben mit ihrem Error-Text + Retry-
  Chip stehen, sodass der User korrigieren oder nochmal abschicken
  kann ohne neu zu tippen.
- Credit-Total wird vor Submit angezeigt (Quality × filledRows).
  Story-Room (MAX_PANELS_PER_STORY − panelCount) clampt die
  sichtbaren Zeilen — Batch wird abgelehnt wenn die Story am
  12-Panel-Hard-Cap ist.
- Shared style (Stilprefix) + character refs werden identisch zum
  Einzel-Flow aus der Story gezogen; jede Batch-Zeile geht durch
  den gleichen `runPanelGenerate`-Pfad wie M2, also kein
  Divergenz-Risiko.

Kein AI-Storyboard (M4), keine MCP-Tools (M5).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:59:17 +02:00
Till JS
bfa923dc22 feat(writing): M5 — cross-module references in the briefing
Drafts can now pull in saved articles, notes, library entries, and raw
URLs as prompt context. This is the Writing module's main differentiator
against standalone LLM chat: the user's own SSOT flows straight into the
ghostwriter without copy-paste.

- utils/reference-resolver.ts: resolveReference() per kind (article,
  note, library, url) via scopedGet + decryptRecords + module type
  converter. Each ref truncates to MAX_CHARS_PER_REF=1500 (with a
  "[… gekürzt …]" marker); resolveReferences() caps the aggregate at
  MAX_TOTAL_REFERENCE_CHARS=8000 and drops extras rather than slicing
  mid-sentence. Deleted or missing refs silently fall out.
- prompt-builder: buildDraftPrompt() takes resolvedReferences and
  renders them as a "--- Quellen ---" block in the user message with
  [Quelle N] headers + optional "Kontext:" lines (the user's own
  per-ref note). System prompt gets a sentence instructing the model
  to paraphrase from the sources and not fabricate facts when a source
  has nothing useful.
- generations store: startDraftGeneration resolves references in
  parallel before building the prompt. No changes to the refineSelection
  path — M5 keeps selection-refinement context-free on purpose.
- UI: ReferencePicker.svelte inline in the BriefingForm with four kind
  tabs (Artikel / Notiz / Library / URL). Searchable lists per kind for
  module refs (max 20 visible, debounced); URL kind takes a url + an
  optional context note. ReferenceChip.svelte pills render live-
  resolved titles; parent resolves labels via the module queries. Hard
  cap at 6 references per draft.
- Scope limits: kontext / goal / me-image refs are on the roadmap but
  deliberately skipped in M5 — they require different resolution paths
  (singletons, structured metadata, image descriptors) that would
  sprawl this commit.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:53:35 +02:00
Till JS
8c43c119ea fix(calendar): promote visibility picker to a labeled prop-row
User feedback: the picker tucked behind the compact icon in the title
row was easy to miss. The other prop-rows carry an icon + field so a
full row reads as "this is an editable property".

Moves the picker into its own prop-row with a "Sichtbarkeit" label,
matching the pattern todo + places already use. Uses the non-compact
variant so the current level (lock / users / link / globe + word) is
readable at a glance.

prop-row--labeled modifier carries the label-text style (muted, fixed
5rem min-width so the picker aligns with the other rows).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:52:59 +02:00
Till JS
98a68afc74 fix(calendar): visibility picker also in the Workbench DetailView
User reported the picker was missing in the Workbench's inline calendar
detail view. I'd only patched EventDetailModal.svelte (used by the
/calendar route), not views/DetailView.svelte which the app-registry
loads for the Workbench card overlay.

Same pattern as the other fixes: import VisibilityPicker + type,
handleVisibilityChange → eventsStore.setVisibility, place the compact
picker in the title-row next to the title input.

Added a small .title-row flex style so title input and picker sit
side-by-side without overlap. `:global(.title-input)` used because the
class is shared with other DetailViews' overrides.

Lesson: any module with BOTH a route-level detail component AND a
views/DetailView.svelte registered for the Workbench needs the picker
in both. Checked other shipped modules:
  - todo: views/DetailView already had it (correct)
  - places: views/DetailView already had it (correct)
  - library: opens in a sub-route, not inline Workbench — views/
    DetailView covers both (correct)
  - wardrobe: opens in a sub-route /wardrobe/outfit/[id] — covers both
  - picture: no Workbench detail view registered, only list
  - goals/recipes: inline-on-card, no separate detail view

Only calendar had the split, now fixed.

Verified:
- pnpm check (web): 7515 files, 0 errors

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:49:48 +02:00
Till JS
3551652612 feat(comic): M2 — UI + Single-Panel-Generierung
Die Datenschicht aus M1 wird jetzt durch UI + gpt-image-2-Flow
benutzbar. Nutzer legt eine Story an (Titel, Stil, Protagonist) und
generiert Panels einzeln über PanelEditor — jeder Panel-Call nutzt
die story-weite Referenz-Liste (face + optional body + optional
Kostüme) plus den stil-spezifischen Prompt-Prefix aus styles.ts.

- `api/generate-panel.ts` → `runPanelGenerate()` wrappt
  `/picture/generate-with-reference` analog zu wardrobe/try-on,
  schreibt picture.images mit `comicStoryId` + `comicPanelIndex`
  Back-Refs und appendet via `comicStoriesStore.appendPanel`. Größe
  defaultet auf 1024×1024 (Quadrat) bzw. 1024×1536 für Webtoon.
- Form-Komponenten: `StylePicker` (5 Presets als Radio-Tiles),
  `CharacterPicker` (face-ref Pflicht, body-ref + bis 3
  Wardrobe-Kostüme optional), `StoryForm` (Titel + Stil + Picker +
  optionaler Kontext).
- Panel-Komponenten: `PanelCard` (Bild + Caption/Dialog-Sidecar),
  `PanelStrip` (responsives Grid 2-4 Spalten), `PanelEditor`
  (inline-Sheet mit Prompt + Caption + Dialog + Quality/Format +
  Generate-Button; zeigt Credits vorher, warnt ab 8 Panels, cappt
  bei 12).
- `StoryCard` rendert Cover aus `panelImageIds[0]` via neuer
  `usePanelImage`-Query, mit Style-Badge und Favorit-Heart.
- `ListView`: Grid + "+ Neue Story"-CTA, Face-Ref-Hinweis wenn
  fehlt, leeres Empty-State-Board.
- `DetailView`: Meta-Card mit VisibilityPicker + Favorit +
  Archive/Delete, PanelStrip, "+ Panel"-CTA öffnet PanelEditor
  inline. Panel-Remove entfernt aus panelImageIds + panelMeta, die
  picture.images-Row bleibt (Final-Delete im Picture-Modul).
- Routes: `/comic` (ListView), `/comic/new` (StoryForm) und
  `/comic/[id]` (DetailView mit {#key id} Re-Mount wie wardrobe).
- i18n: comic-Label in de.json + en.json für RoutePage-Header.
- queries: `usePanelImage(id)` Helper für Cover + Panel-Rendering
  (comic-intern, nicht ins Picture-Modul eingemischt).

Sprechblasen/Captions werden gpt-image-2 per Prompt übergeben und
direkt ins Bild gerendert — kein SVG-Overlay. Englische Texte
rendern stabiler (UI-Hinweis).

Testet per `pnpm run check` + `validate:all` sauber, 5 Encryption-
Tests weiterhin grün.

Kein Batch-Mode (M3), kein AI-Storyboard (M4), keine MCP-Tools (M5).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:42:27 +02:00
Till JS
19e0f33665 feat(writing): M6 — selection-refinement tools
Users can now select a passage in the editor and refine it in place via
five targeted operations instead of re-generating the whole draft:
Shorten, Expand, Change tone, Rewrite (freeform instruction), Translate.

- Five selection-specific prompt builders in utils/prompt-builder.ts.
  Each forbids preamble / quoting / explanation so the output is a drop-
  in replacement for the selected text. Style context is injected when
  present so refinements stay on-voice.
- generations.store.refineSelection() sizes the token budget to the
  selection (selectionWords * 4 + 200), runs at temperature 0.4 for
  consistency, and records the attempt as a LocalGeneration with kind
  'selection-*' + inputSelection range regardless of whether the user
  accepts — every refine-attempt stays auditable.
- applyRefinement() commits the replacement to the current version's
  content (not a new version; in-place per the plan) and back-links the
  generation via outputVersionId so later audits can trace each edit.
- SelectionToolbar appears above the editor when the user has a non-
  empty selection; Tone + Translate expand to pickers, Rewrite to a
  text input.
- RefinementPanel shows original + refined side-by-side with Übernehmen /
  Noch mal / Verwerfen. Running and failed states get their own chrome.
- VersionEditor tracks textarea selection via select/mouseup/keyup and
  reports {start, end, text} via onselect. New `forceContent` prop
  nonce lets the parent swap the editor's local text after an apply or
  undo without breaking the debouncing layer.
- One-step undo: "↶ Rückgängig: <tool>" button surfaces briefly after
  an accepted refinement and restores the pre-refinement content via
  draftsStore.updateVersionContent. Kicking off a new refinement
  clears the undo target.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:37:04 +02:00
Till JS
27c1860f82 feat(comic): M1 — Datenschicht + Modul-Registrierung
Neues Comic-Modul: aus Text-Inputs (Journal / Notes / Writing / Library
/ Calendar) entsteht ein mehrseitiger Comic, generiert mit gpt-image-2
über die bestehende /picture/generate-with-reference-Route. Plan in
docs/plans/comic-module.md (M1–M5 + optional M6–M8).

M1 schafft die Datenschicht ohne UI:
- Dexie v44 `comicStories` (space-scoped, Indices createdAt/style/
  isFavorite/isArchived). Story hält `panelImageIds: string[]` und
  `panelMeta: Record<panelImageId, {caption, dialogue, promptUsed,
  sourceInput?}>` — Panels selbst sind picture.images-Rows mit
  comicStoryId + comicPanelIndex Back-Refs.
- Fünf Stil-Presets (comic / manga / cartoon / graphic-novel / webtoon)
  mit Prompt-Prefix-Templates in styles.ts; composePanelPrompt webt
  Stil + Panel-Prompt + Caption + Dialog zusammen. Sprechblasen
  werden von gpt-image-2 direkt ins Bild gerendert — kein SVG-Overlay.
- Encryption-Registry-Eintrag: title / description / storyContext /
  tags / panelMeta als JSON-Blob. Struktur (id, style, character-
  MediaIds, panelImageIds, Flags, visibility) bleibt plaintext.
- Module-Registry registriert appId='comic', verifyMediaOwnership auf
  der /picture/generate-with-reference-Route akzeptiert jetzt
  ['me', 'wardrobe', 'comic'] — 'comic'-Slot ist reserviert für M6+
  Anchor-/Backdrop-Uploads.
- Space-Allowlist: comic in brand (Marken-Storys), club (Vereins-
  geschichte), family (Kinder-Abenteuer), team (Release-Comics),
  practice (Patienten-Aufklärung). Personal via '*'-Sentinel.
- mana-apps.ts Eintrag mit comic-Icon (Sprechblase + Lightning-Bolt,
  f97316→dc2626 Gradient). Lokal tier='guest' mit LOCAL TIER PATCH-
  Comment wie Wardrobe, canonical ist 'beta'.

Visibility-System von Anfang an adopted (setVisibility-Methode im
Store, unlistedToken-Generierung inklusive). appendPanel() als
Vorarbeit für M2 bereits da, ohne Aufrufer.

5 Encryption-Roundtrip-Tests grün (panelMeta nested JSON, leeres
panelMeta, partielle panelMeta ohne sourceInput, null-description).
pnpm run check + validate:all sauber (207 Dexie-Tabellen klassifiziert,
comicStories unter den 106 encrypted).

Kein UI, keine Panel-Generierung, keine MCP-Tools — alles M2/M3/M5.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:29:51 +02:00
Till JS
1c82a374fe feat(writing): M4 — style system with presets + custom styles
Styles are selectable in the briefing and flow into the prompt builder
so "LinkedIn-Post" and "Akademisch" produce visibly different drafts
from the same brief.

- StylePicker.svelte: dropdown in BriefingForm grouped into Vorlagen
  (9 presets from presets/styles.ts) and Meine Stile (user custom rows).
  Emits an opaque id — `preset:<id>` for presets or a uuid for customs —
  so selecting a preset requires no Dexie write.
- generations store: loadStyle() now resolves both prefix shapes. The
  prompt builder already honoured both preset.principles and
  row.extractedPrinciples, so no prompt changes needed.
- /writing/styles view: grid of presets (read-only dashed cards) plus
  a user section with create / edit / delete for custom styles.
- StyleForm.svelte: M4 supports source='custom-description' (name +
  freeform prose the LLM reads verbatim). Sample-trained and
  self-trained sources come in M4.1.
- DetailView surfaces the active style as a 🎨-chip next to the
  briefing preview; ListView gets a "🎨 Stile" link to the management
  route.

Styles are optional — existing drafts with styleId=null keep their
previous behaviour, and the LinkedIn/Hemingway/etc. presets are a zero-
friction on-ramp before users bother writing a custom one.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:20:06 +02:00
Till JS
d725a8df8b feat(writing): M3 — one-shot prose generation via mana-llm
Server:
- New llmText() helper in apps/api/src/lib/llm.ts for plain-text
  (non-streaming) completions with token-usage reporting.
- POST /api/v1/writing/generations (Hono + requireTier('beta'))
  accepts system+user prompts, forwards to mana-llm (default model
  ollama/gemma3:4b), returns raw output + model + tokenUsage. The
  endpoint is stateless — draft/version bookkeeping is entirely
  client-side so the same route serves refinement calls later.

Client:
- writing/api.ts — Bearer-authed fetch client (follows the food/
  news-research pattern).
- writing/utils/prompt-builder.ts — pure builder turning a briefing
  (+ optional style preset / extracted principles) into a system+user
  pair. Forbids preamble / sign-off / meta commentary so the output is
  ready to paste into a version.
- writing/stores/generations.svelte.ts — orchestrates the full flow:
  queued → running → call → new LocalDraftVersion → pointer flip →
  succeeded. On failure leaves the current version untouched with the
  error on the generation record. Emits WritingDraftGenerationStarted /
  WritingDraftVersionCreated / WritingDraftGenerationFailed events.

UI:
- Generate button in DetailView.svelte (label flips "Generate" / "Neu
  generieren" based on whether the draft already has content).
- GenerationStatus.svelte strip surfaces queued / running / failed with
  model + duration badges; succeeded generations auto-disappear because
  the new version is already live via the currentVersionId pointer.

M3 is synchronous and non-streaming by design. M7 adds mission-based
long-form with streaming + outline stage + reference injection. M6 will
reuse the same /generations endpoint for selection-refinement prompts.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 15:11:48 +02:00