feat(config): rewire model roles to frank-small/frank-large + NPU flm provider

- models.yml: rename frank-default-small -> frank-small (dankserver),
  frank-default-large -> frank-large (dankclaw-gpu, text-only input);
  add 'flm (dankclaw-npu)' provider with gemma4-it:e4b and nuextract3:4b.
- config.yml: point modelRoles at renamed models; vision -> nuextract3:4b:auto,
  tiny/task -> gemma4-it:e4b:off; compaction reorder (snapcompact added,
  idleThresholdTokens 200k->100k, thresholdPercent -1->85); hideThinkingBlock true;
  statusline transparency + compact thinking level; images.urls backends list.

agent/config.yml

@@ -24,16 +24,16 @@ providers:
     - mojeek
     - public
 modelRoles:
-  default: llamacpp (dankclaw)/frank-default-large:medium
+  smol: llamacpp (dankserver)/frank-small:off
-  smol: llamacpp (dankserver)/frank-default-small:xhigh
+  slow: llamacpp (dankclaw)/frank-large:medium
-  slow: llamacpp (dankclaw)/frank-default-large:medium
+  vision: flm (dankclaw-npu)/nuextract3:4b:auto
-  vision: llamacpp (dankclaw)/frank-default-large:xhigh
+  plan: llamacpp (dankclaw)/frank-large:medium
-  plan: llamacpp (dankclaw)/frank-default-large:medium
+  designer: llamacpp (dankclaw)/frank-large:medium
-  designer: llamacpp (dankclaw)/frank-default-large:medium
+  commit: llamacpp (dankserver)/frank-small:medium
-  commit: llamacpp (dankserver)/frank-default-small
+  tiny: flm (dankclaw-npu)/gemma4-it:e4b:off
-  tiny: llamacpp (dankserver)/frank-default-small
+  task: flm (dankclaw-npu)/gemma4-it:e4b:off
-  task: llamacpp (dankserver)/frank-default-small
+  advisor: llamacpp (dankclaw)/frank-large:medium
-  advisor: llamacpp (dankclaw)/frank-default-large:medium
+  default: llamacpp (dankserver)/frank-small:medium
 symbolPreset: nerd
 composer:
   shape: claude
@@ -57,16 +57,17 @@ autolearn:
 compaction:
   methodOrder:
     - shake
-    - handoff
     - soft
     - remote
-  idleThresholdTokens: 200000
+    - handoff
-  thresholdPercent: -1
+    - snapcompact
+  idleThresholdTokens: 100000
+  thresholdPercent: 85
   idleEnabled: true
   thresholdTokens: 150000
 interruptMode: immediate
 defaultThinkingLevel: medium
-hideThinkingBlock: false
+hideThinkingBlock: true
 proseOnlyThinking: true
 terminal:
   showProgress: true
@@ -75,9 +76,9 @@ disabledExtensions:
 statusLine:
   separator: powerline
   preset: nerd
-  contextLine: annotated
+  contextLine: "off"
-  transparent: false
+  transparent: true
-  compactThinkingLevel: false
+  compactThinkingLevel: true
   showHookStatus: true
 tui:
   textSizing: true
@@ -117,3 +118,14 @@ loop:
   mode: prompt
 steeringMode: one-at-a-time
 personality: default
+images:
+  urls:
+    backends:
+      - provider-files
+      - tailscale
+      - cloudflared
+      - litterbox
+edit:
+  mode: hashline
+dev:
+  autoqaConsent: granted

agent/models.yml

@@ -7,12 +7,12 @@ providers:
     api: openai-completions
     auth: none
     models:
-      - id: frank-default-small
+      - id: frank-small
-        name: "frank-default-small (Qwen3.6-35B-A3B FP8, FreeToken RTX3090, moe-cache 2048)"
+        name: "frank-small (dankserver)"
-        contextWindow: 204800
+        contextWindow: 262144
-        maxTokens: 204800
+        maxTokens: 262144
         reasoning: true
-        input: [text, image]
+        input: [text]
         compat:
           reasoningEffortMap:
             max: xhigh
@@ -22,12 +22,34 @@ providers:
     api: openai-completions
     auth: none
     models:
-      - id: frank-default-large
+      - id: frank-large
-        name: "frank-default-large (Qwen3.8-Flash-Next Q4_0 ROCmFP4 Strix, froggeric-v22.4.0 template)"
+        name: "frank-large (dankclaw-gpu)"
-        contextWindow: 262144
+        contextWindow: 200000
-        maxTokens: 262144
+        maxTokens: 200000
+        reasoning: true
+        input: [text]
+        compat:
+          reasoningEffortMap:
+            max: xhigh
+
+  "flm (dankclaw-npu)":
+    baseUrl: http://192.168.0.143:7771/v1
+    api: openai-completions
+    auth: none
+    models:
+      - id: gemma4-it:e4b
+        name: "gemma4-it:e4b (dankclaw-npu)"
+        contextWindow: 131072
+        maxTokens: 131072
         reasoning: true
         input: [text, image]
         compat:
           reasoningEffortMap:
             max: xhigh
+
+      - id: nuextract3:4b
+        name: "nuextract3:4b (dankclaw-npu extraction)"
+        contextWindow: 32768
+        maxTokens: 32768
+        reasoning: false
+        input: [text]