Initial TrueGrowth source import

openspec/changes/add-digital-human-script-batch/proposal.md

@@ -0,0 +1,17 @@
+# Change: Add digital human script batch generation
+
+## Why
+
+Digital human creation currently supports one script at a time. Operators need to generate or add multiple talking-head scripts, then create several digital human videos from the same avatar, voice, and model configuration without repeating the form.
+
+## What Changes
+
+- Add a script list to the digital human one-click compose flow, supporting manual single-script addition and multi-script batch editing.
+- Add an intelligent script generation dialog with built-in industry-style digital human talking-head scenes, word/time/style/count controls, and custom prompt input.
+- Append generated script results into the script list instead of replacing existing work.
+- Allow batch creation of digital human tasks, creating one runtime task per selected script while reusing the selected avatar, voice, model, subtitle, format, and background settings.
+
+## Impact
+
+- Affected specs: digital-human-workflow
+- Affected code: `apps/web/src/workbench/WorkbenchShell.tsx`, `apps/web/src/workbench/workbench.scss`

openspec/changes/add-digital-human-script-batch/specs/digital-human-workflow/spec.md

@@ -0,0 +1,31 @@
+## ADDED Requirements
+
+### Requirement: Digital Human Script List
+
+The system SHALL let users maintain a list of one or more digital human talking-head scripts inside the one-click compose flow before creating videos.
+
+#### Scenario: Add scripts manually
+
+- **WHEN** a user enters a script and clicks add
+- **THEN** the script is appended to the current script list
+- **AND** the user can select, edit, duplicate, or remove scripts before generation
+
+### Requirement: Smart Digital Human Script Generation
+
+The system SHALL provide an intelligent script generation dialog with built-in digital human scenes and controls for count, length, style, and custom prompt guidance.
+
+#### Scenario: Generate scripts from a scene
+
+- **WHEN** a user chooses a scene, target length, style, count, and optional custom prompt
+- **THEN** the system generates the requested number of talking-head scripts
+- **AND** appends the generated scripts to the current script list without discarding existing scripts
+
+### Requirement: Batch Digital Human Creation
+
+The system SHALL create a separate digital human generation task for each selected script while reusing the chosen avatar, voice, model, subtitle, aspect ratio, and background settings.
+
+#### Scenario: Start batch generation
+
+- **WHEN** the user selects multiple scripts and clicks batch create
+- **THEN** one digital human task is created for each selected script
+- **AND** every task payload uses the corresponding script text with the current avatar and voice configuration

openspec/changes/add-digital-human-script-batch/tasks.md

@@ -0,0 +1,7 @@
+## 1. Implementation
+
+- [x] 1.1 Add digital human script item state, manual add/edit/delete/selection controls, and single/multi-script count display.
+- [x] 1.2 Add smart script generation dialog with built-in scenes, style/length/count/custom prompt controls, and append-to-list behavior.
+- [x] 1.3 Update avatar task creation to submit one task per selected script while preserving the selected avatar, voice, model, subtitle, format, and background settings.
+- [x] 1.4 Style the batch script list and script generation dialog for desktop and responsive layouts.
+- [x] 1.5 Run focused type/build validation. `tsc -p apps/web/tsconfig.app.json --noEmit` passes; full `nx build web` is blocked by existing `.venv-funclip` Gradio project graph noise.