feat: workspaces shell + cwd-lsp rename + mcp/settings/system-prompt features + app wiring

- workspaces: URL-driven conversation grouping (home listing at /, routing,
  store, http adapter, WorkspaceCard) wired into the App.svelte shell
- rename features/workspace -> features/cwd-lsp (the cwd/lsp status feature)
- new features: mcp (status view), settings (chat-limit field), system-prompt
  (prompt builder), all rendered via the generic surface host
- chat: store + ChatView updates
- tabs: tabs-store updates
- app wiring: ErrorModal (full-screen error surface), app/App.svelte + store.svelte

This commit makes HEAD typecheck clean for the first time: the prior HEAD
(dfe1711) imported features/settings from app/App.svelte but never committed
the feature, so only the full working tree was green.

1 files changed, 60 insertions, 0 deletions

diff --git a/src/features/settings/logic/view-model.ts b/src/features/settings/logic/view-model.ts
new file mode 100644
index 0000000..caf76ee
--- /dev/null
+++ b/src/features/settings/logic/view-model.ts
@@ -0,0 +1,60 @@
+import { normalizeChatLimit } from "../../../core/chunks";
+
+/**
+ * Pure core for the settings feature — zero DOM, zero effects, zero Svelte.
+ *
+ * The settings feature exposes FE-local, user-tunable settings. Currently the
+ * chat limit (max loaded chunks per conversation; the policy lives in
+ * `core/chunks/trim.ts`): this module is the view-model seam that parses +
+ * validates a typed value into a normalized limit, so the UI can disable submit
+ * and show an error on garbage input. The bound constants + normalization are
+ * OWNED by `core/chunks` (the single source of truth); this module never
+ * redefines them.
+ */
+
+// ── Injected port (consumer-defines-port; the composition root adapts the
+//    store's localStorage persistence to this shape). ──────────────────────────
+
+/** Outcome of persisting a chat-limit setting. */
+export type ChatLimitSaveResult =
+	| { readonly ok: true; readonly chatLimit: number }
+	| { readonly ok: false; readonly error: string };
+
+export type SaveChatLimit = (value: number) => Promise<ChatLimitSaveResult>;
+
+// ── chat-limit parse / validate ───────────────────────────────────────────────
+
+/** Result of parsing a typed chat-limit string. */
+export type ChatLimitParse =
+	| { readonly ok: true; readonly value: number }
+	| { readonly ok: false; readonly error: string };
+
+/**
+ * Parse a typed chat-limit string into a normalized limit (floored + clamped to
+ * [MIN_CHAT_LIMIT, MAX_CHAT_LIMIT] via `normalizeChatLimit`). Empty or
+ * non-numeric input is an ERROR (so the UI can disable submit + message), NOT a
+ * silent default — the default only applies at boot for an unset key, never
+ * from user typing.
+ */
+export function parseChatLimit(raw: string): ChatLimitParse {
+	const trimmed = raw.trim();
+	if (trimmed.length === 0) {
+		return { ok: false, error: "Enter a number." };
+	}
+	const n = Number(trimmed);
+	if (!Number.isFinite(n)) {
+		return { ok: false, error: "Must be a number." };
+	}
+	return { ok: true, value: normalizeChatLimit(n) };
+}
+
+/**
+ * Whether saving `typed` would change the `current` chat limit. A no-op save
+ * (empty/invalid, or equal after normalization) should be disabled. This is the
+ * dirty-check for the input — it must NOT mutate or clamp, only compare.
+ */
+export function chatLimitChanged(typed: string, current: number): boolean {
+	const parsed = parseChatLimit(typed);
+	if (!parsed.ok) return false;
+	return parsed.value !== current;
+}