chore(release): v0.6.0 - Gemma 4 thinking ON usable + schema-driven env UI

Highlights
- thinking ON for Gemma 4 26B-A4B IT is production-usable via two
  safety net layers: ChannelBudget (force-close at N reasoning tokens)
  and n-gram runaway detector. Validated end-to-end with Ayla against
  the mlx-community baseline and our hybrid embed8-last13attn8 ship.
- LUMEN_BACKEND_THINKING_DEFAULT env restored. Operator opt-in for
  per-server reasoning default; per-request signals
  (chat_template_kwargs.enable_thinking, reasoning_effort, thinking)
  still override.
- Schema-driven env overrides UI in lumen-app: 12 typed knobs
  (Reasoning / Safety net / Advanced / Debug), checkbox/number-with-
  range/select widgets, validated input, bilingual labels.
- Embedding visibility fix: scan_local now merges entries from the
  hf-hub cache, so models auto-fetched by the server appear as
  downloaded in the UI. New MODELS-card embedding mini-row.
- Log viewer: auto-scroll to bottom, content-visibility virtualization,
  raised in-memory cap from 500 to 2000 lines.

mlx-rs / mlx-sys git revs unchanged from v0.5.0.

Recommended Gemma 4 server env for thinking ON:
  LUMEN_BACKEND_THINKING_DEFAULT=on
  LUMEN_MAX_THINKING_TOKENS=600
  LUMEN_MAX_FORCE_CLOSE_ATTEMPTS=1
  LUMEN_RUNAWAY_DETECT=1

Author: rabbitson87

Cargo.lock

@@ -1,6 +1,6 @@
 [package]
 name = "lumen-app"
-version = "0.5.0"
+version = "0.6.0"
 edition.workspace = true
 rust-version.workspace = true
 license.workspace = true

crates/lumen-app/Cargo.toml

@@ -1,7 +1,7 @@
 {
   "name": "lumen-app-frontend",
   "private": true,
-  "version": "0.5.0",
+  "version": "0.6.0",
   "type": "module",
   "scripts": {
     "predev": "cargo build --manifest-path ../../../Cargo.toml -p lumen-server --release",

crates/lumen-app/frontend/package.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://schema.tauri.app/config/2",
   "productName": "Lumen",
-  "version": "0.5.0",
+  "version": "0.6.0",
   "identifier": "ai.lumen.app",
   "build": {
     "frontendDist": "frontend/dist",