catalog: re-source today's engine defaults from canonical upstreams

Replace the invented sample inputs / topk with canonically-sourced values
(per CATALOG-CONTRACT source precedence + cite-your-source rule), inline-cited:
  * dia2-2b/dia2-1b input -> '[S1] Hello Dia2!'  (nari-labs/dia2 README example)
  * zonos input -> 'Zonos uses eSpeak for text to phoneme conversion!'
      (upstream gradio_interface.py blessed Textbox default)
  * csm input -> 'Hello from Sesame.'  (sesame/csm-1b model card example)
  * csm topk -> default REMOVED (the card recommends no top-k default; the
    wrapper's internal default applies when omitted).

docs/asset-engine/services.yaml

@@ -1268,7 +1268,7 @@ services:
         label: Text ([S1]/[S2] dialogue + nonverbals)
         required: true
         max_length: 5000
-        default: "[S1] Did it actually work? [S2] (laughs) Worked on the first try."
+        default: "[S1] Hello Dia2!"  # canonical: nari-labs/dia2 README usage example
         description: >
           [S1]/[S2] tags mark speaker turns; nonverbals like (laughs),
           (coughs), (sighs), (clears throat) go inline.
@@ -1347,7 +1347,7 @@ services:
         label: Text ([S1]/[S2] dialogue + nonverbals)
         required: true
         max_length: 5000
-        default: "[S1] Did it actually work? [S2] (laughs) Worked on the first try."
+        default: "[S1] Hello Dia2!"  # canonical: nari-labs/dia2 README usage example
         description: >
           [S1]/[S2] tags mark speaker turns; nonverbals like (laughs),
           (coughs), (sighs), (clears throat) go inline.
@@ -1426,7 +1426,7 @@ services:
         label: Text
         required: true
         max_length: 5000
-        default: "Hey, good to finally talk with you. How has your day been?"
+        default: "Hello from Sesame."  # canonical: sesame/csm-1b model card
       - name: voice
         type: select
         label: Voice
@@ -1453,8 +1453,7 @@ services:
       - name: topk
         type: number
         required: false
-        default: 50
-        description: Top-k sampling cutoff (1–100).
+        description: Top-k sampling cutoff (1–100); no canonical default — wrapper's internal default applies if omitted.
       - name: max_audio_length_ms
         type: number
         required: false
@@ -1530,7 +1529,7 @@ services:
         section: text
         required: true
         max_length: 500
-        default: "This is a quick test of an expressive, multilingual voice."
+        default: "Zonos uses eSpeak for text to phoneme conversion!"  # canonical: upstream gradio_interface.py default
         description: eSpeak-phonemized. Upstream Gradio caps the box at ~500 chars.
       - name: language
         type: select