whiteport-design-studio 0.2.2 → 0.3.0

package/src/workflows/4-ux-design/steps-s/step-15-create-page-structure.md

@@ -17,7 +17,7 @@ Create the physical page folder structure, generate the initial specification do
 
 ### Universal Rules:
 
-- 🛑 NEVER generate content without user input
+- 🛑 NEVER generate content without user input (Suggest mode) / Generate based on context and WDS patterns (Dream mode)
 - 📖 CRITICAL: Read the complete step file before taking any action
 - 🔄 CRITICAL: When loading next step with 'C', ensure entire file is read
 - 📋 YOU ARE A FACILITATOR, not a content generator
@@ -60,13 +60,13 @@ Create the physical page folder structure, generate the initial specification do
 **Determine page folder path:**
 
 **For single page projects (no scenarios):**
-- Page path: `4-scenarios/{{page_slug}}/`
+- Page path: `C-UX-Scenarios/{{page_slug}}/`
 
 **For scenario-based projects:**
 - Read scenario_number from context
 - Read current_page_index from `scenario-tracking.yaml`
 - Calculate page_number: `{{scenario_number}}.{{current_page_index + 1}}`
-- Page path: `4-scenarios/{{scenario_number}}-{{scenario-slug}}/{{page_number}}-{{page_slug}}/`
+- Page path: `C-UX-Scenarios/{{scenario_number}}-{{scenario-slug}}/{{page_number}}-{{page_slug}}/`
 
 Store page_path and page_number
 </action>
@@ -178,25 +178,44 @@ If this is a scenario-based project:
 - Add sketches to the sketches folder
 - Continue with page design</output>
 
-### 2. Present MENU OPTIONS
+### 2. Two-Option Transition
 
-Display: "**Select an Option:** [A] Add Another Page | [S] Add Another Scenario | [M] Return to Activity Menu"
+After page structure is created, present exactly two options:
 
-#### Menu Handling Logic:
+**If more pages exist in the scenario (from Q8 shortest path):**
 
-- IF A: Load, read entire file, then execute ./step-08-page-context.md
-- IF S: Load, read entire file, then execute ./step-01-core-feature.md
-- IF M: Return to {workflowFile} or {activityWorkflowFile}
-- IF Any other comments or queries: help user respond then [Redisplay Menu Options](#2-present-menu-options)
+<output>
+**Page structure for "[page name]" is ready!**
+
+1. **Specify this page** — add full detail with [P] Specify
+2. **Design the next scenario step** — [next page name]
+</output>
+
+**If this is the last page in the scenario:**
+
+<output>
+**Page structure for "[page name]" is ready!**
+
+1. **Specify this page** — add full detail with [P] Specify
+2. **All pages in this scenario are created!** — return to dashboard
+</output>
+
+#### Transition Handling:
+
+- **Option 1 (specify):** Load and execute `../steps-p/step-01-page-basics.md`
+- **Option 2 (next page):** Load and execute `./step-08-page-context.md` for the next page
+- **Option 2 (all done):** Return to {workflowFile} adaptive dashboard
+- **Dream mode:** Auto-proceed to option 1. Skip menu display.
 
 #### EXECUTION RULES:
 
-- ALWAYS halt and wait for user input after presenting menu
-- User can chat or ask questions — always respond and then redisplay menu options
+- **Suggest mode:** ALWAYS halt and wait for user input after presenting transition options
+- User can chat or ask questions — always respond and then redisplay the transition
+- The user can always say "stop" to pause and return later — log current status
 
 ## CRITICAL STEP COMPLETION NOTE
 
-ONLY WHEN the user selects an option from the menu and the page structure has been created will you proceed as directed. This is the last step in the Suggest activity chain.
+ONLY WHEN the user selects an option and the page structure has been created will you proceed as directed. This is the last step in the Suggest page creation chain.
 
 ---
 

package/src/workflows/4-ux-design/steps-v/step-03-page-overview.md

@@ -63,7 +63,7 @@ Validate:
 - Page description paragraph (1-2 paragraphs explaining what page does)
 - User Situation section (user's context, needs, emotional state)
 - Page Purpose section (what job page must accomplish)
-- Success criteria or VTC reference
+- Success criteria or Trigger Map reference
 - Emotional context and pain points addressed
 
 ### 2. Generate Diagnostic Report

package/src/workflows/4-ux-design/steps-w/step-01-visual-approach.md

@@ -82,8 +82,7 @@ How would you like to create the visual design?
 ### 3. Setup Gate (Nano Banana only)
 
 If user selects [N]:
-1. Check for existing agent dialog for this page in `{output_folder}/_progress/agent-dialogs/`
-   - If active dialog found: offer to resume previous visual generation session
+1. Check the design log at `{output_folder}/_progress/00-design-log.md` for previous visual generation entries for this page
 2. If first time using Nano Banana in this project:
    - Route to `step-00-nb-setup.md` to verify MCP connection
    - Return here after verification succeeds

package/src/workflows/4-ux-design/steps-w/step-02-generate-visual.md

@@ -61,7 +61,7 @@ Create the visual design using the chosen tool — route to the appropriate sub-
 <action>Load and execute: step-02w-nb-compose-prompt.md</action>
 
 This sub-workflow handles:
-- Agent dialog creation (tracks prompts and generation history)
+- Design log entry (tracks prompts and generation history)
 - Image description extraction from the page spec
 - User creative direction (overrides and enhancements)
 - Prompt composition with compression strategy

package/src/workflows/4-ux-design/steps-w/step-02w-nb-compose-prompt.md

@@ -39,12 +39,12 @@ Translate a page specification into an effective AI image generation prompt that
 - 🎯 Focus on composing an effective generation prompt from page spec data
 - 🚫 FORBIDDEN to generate without user confirming creative direction
 - 💬 Approach: Extract, present, let user override, compose, generate, iterate
-- 📋 Track all generations in agent dialog file
+- 📋 Track all generations in agent experience file
 
 ## EXECUTION PROTOCOLS:
 
 - 🎯 Extract image descriptions, gather creative direction, compose prompt, generate
-- 💾 Log all generations to agent dialog file
+- 💾 Log all generations to agent experience file
 - 📖 Reference NANO-BANANA-PROMPT-GUIDE.md for compression strategy
 - 🚫 FORBIDDEN to skip creative direction step
 
@@ -57,24 +57,15 @@ Translate a page specification into an effective AI image generation prompt that
 
 ## Sequence of Instructions (Do not deviate, skip, or optimize)
 
-### 0. Start Agent Dialog
+### 0. Start Generation Log
 
-Create (or resume) an agent dialog file to track this visual generation session.
+Create an agent experience file to track this visual generation session.
 
-**File location:** `{output_folder}/_progress/agent-dialogs/`
-**Naming:** `{date}-freya-visual-{page-name}.md`
-**Example:** `2026-02-19-freya-visual-1.1-hem.md`
+**File location:** `{output_folder}/_progress/agent-experiences/`
+**Naming:** `{date}-visual-{page-name}.md`
+**Example:** `2026-02-19-visual-1.1-hem.md`
 
 ```markdown
----
-status: active
-agent: Freya
-topic: Visual generation for {page-name}
-created: {timestamp}
-last_updated: {timestamp}
-context_summary: Generating visual designs for {page-name} using Nano Banana
----
-
 # Visual Generation: {page-name}
 
 ## Inputs
@@ -92,7 +83,7 @@ context_summary: Generating visual designs for {page-name} using Nano Banana
 {each generation appended here in step H}
 ```
 
-**If an active dialog exists** for this page, resume it -- load previous context, creative direction, and generation history.
+**If a previous generation log exists** for this page, read it for context — previous creative direction, successful prompts, and lessons learned.
 
 ### A. Load Inputs
 
@@ -253,7 +244,7 @@ Call `mcp__nanobanana__generate_image` with the assembled prompt, system instruc
 
 Present the result to the user.
 
-**Log to agent dialog:**
+**Log to agent experience file:**
 
 ```markdown
 ### Generation {N} -- {timestamp}
@@ -274,7 +265,7 @@ Present the result to the user.
 **User feedback:** {filled after review}
 ```
 
-Update the dialog `last_updated` timestamp.
+Update the agent experience file.
 
 ### I. Iterate
 
@@ -339,19 +330,19 @@ ONLY WHEN the user has accepted a generated visual and selected an option from t
 
 ### ✅ SUCCESS:
 
-- Agent dialog created or resumed
+- Generation log created in agent experiences
 - Image descriptions extracted from spec
 - User creative direction captured
 - Prompt composed within 8192 char limit
 - Image generated and presented
-- Generation logged to agent dialog
+- Generation logged to agent experience file
 - User accepted or iterated to satisfaction
 
 ### ❌ SYSTEM FAILURE:
 
 - Generating without user creative direction
 - Exceeding prompt character limit
-- Not logging generations to agent dialog
+- Not logging generations to agent experience file
 - Not presenting iteration options
 - Skipping reference image auto-detection
 

package/src/workflows/4-ux-design/steps-w/step-03-review-integrate.md

@@ -85,9 +85,9 @@ Store visual artifact in the appropriate location:
 
 - **UI mockups (page/section):** `{output_folder}/D-Design-System/01-Visual-Design/design-concepts/`
 - **Image assets (photos/illustrations):** `{output_folder}/D-Design-System/01-Visual-Design/design-concepts/` (move to `02-Assets/images/` when finalized)
-- **Legacy path:** `{output_folder}/D-UX-Design/[scenario]/visuals/` (if project uses older folder structure)
+- **Legacy path:** `{output_folder}/C-UX-Scenarios/[scenario]/visuals/` (if project uses older folder structure)
 
-**Update the agent dialog** with the accepted result and save path.
+**Update the agent experience file** with the accepted result and save path.
 
 ### 5. Present MENU OPTIONS
 
@@ -117,14 +117,14 @@ ONLY WHEN the user selects an option from the menu and the visual has been integ
 - Design decisions documented
 - Page specification updated with visual reference
 - Visual artifact saved to correct location
-- Agent dialog updated with accepted result
+- Agent experience file updated with accepted result
 
 ### ❌ SYSTEM FAILURE:
 
 - Skipping user feedback
 - Not integrating into page specification
 - Not saving visual artifact
-- Not updating agent dialog
+- Not updating agent experience file
 - Relying on AI-generated text in images for content accuracy
 
 **Master Rule:** Skipping steps, optimizing sequences, or not following exact instructions is FORBIDDEN and constitutes SYSTEM FAILURE.

package/src/workflows/4-ux-design/templates/audit-report.template.md

@@ -102,7 +102,7 @@
 **Checklist:**
 - [ ] User situation clearly defined
 - [ ] Usage context documented
-- [ ] Trigger Value Chain (VTC) defined and linked
+- [ ] Strategic context (Trigger Map) defined and linked
 - [ ] Scenario purpose stated
 - [ ] Success criteria defined
 
@@ -134,7 +134,7 @@
 **Checklist:**
 - [ ] Page purpose clearly stated
 - [ ] Success criteria defined
-- [ ] VTC reference present
+- [ ] Trigger Map reference present
 - [ ] Sections properly separated and named
 - [ ] Section purposes defined
 - [ ] Page layout logical and flows well

package/src/workflows/4-ux-design/templates/page-specification.template.md

@@ -82,6 +82,49 @@
 
 ---
 
+## Spacing
+
+<!--
+  All spacing values MUST use token names from the project's spacing scale.
+  The scale is defined in D-Design-System/00-design-system.md → Spacing Scale.
+  This can be the WDS default scale, Tailwind classes, Material tokens, or any system.
+  Do NOT use arbitrary pixel values. If unsure, check the scale.
+-->
+
+**Scale:** [Spacing Scale](../../D-Design-System/00-design-system.md#spacing-scale)
+
+| Property | Token |
+|----------|-------|
+| Page padding (horizontal) | {e.g., space-md mobile / space-lg desktop} |
+| Section gap | {e.g., space-xl} |
+| Element gap (default within sections) | {e.g., space-md} |
+| Component gap (within groups) | {e.g., space-sm} |
+
+---
+
+## Typography
+
+<!--
+  Text sizes use token names from the project's type scale.
+  The scale is defined in D-Design-System/00-design-system.md → Type Scale.
+
+  IMPORTANT: Semantic level (H1, H2, p) signals document structure — for accessibility and SEO.
+  Visual styling (size, weight, typeface) signals visual hierarchy — how it looks.
+  They are related but NOT locked together. An H2 can be text-sm. A <p> can be text-2xl.
+  Headings can have different typefaces and styles on different pages — that's fine.
+-->
+
+**Scale:** [Type Scale](../../D-Design-System/00-design-system.md#type-scale)
+
+| Element | Semantic | Size | Weight | Typeface |
+|---------|----------|------|--------|----------|
+| {Page title} | H1 | {e.g., text-2xl} | {e.g., bold} | {e.g., display / default} |
+| {Section heading} | H2 | {e.g., text-xl} | {e.g., semibold} | {default} |
+| {Body text} | p | text-md | normal | {default} |
+| {Caption/helper} | p | {e.g., text-xs} | normal | {default} |
+
+---
+
 ## Page Sections
 
 ### Section: {Section Name}
@@ -92,6 +135,8 @@
 |----------|-------|
 | Purpose | {What this section does} |
 | Component | [{Design System Component}]({path}) |
+| Padding | {e.g., space-lg space-md} |
+| Element gap | {e.g., space-md — or "default" if same as page-level} |
 
 ---
 
@@ -107,6 +152,45 @@
 | EN | "{English text}" |
 | Behavior | {onClick / onChange / etc.} |
 
+#### ↕ `{page}-{v|h}-{type}-{size}` — {reason}
+
+<!--
+  Spacing objects sit between content objects. They have IDs and are first-class.
+
+  NAMING: {page}-{v|h}-{type}-{size}
+  - v = vertical, h = horizontal
+  - type = space, separator, line
+  - size = the token name (zero, sm, md, lg, xl, 2xl, 3xl, flex)
+  The ID describes WHAT the spacing IS, not which objects it sits between.
+
+  RULES:
+  - Default element spacing (from the Spacing section above) is implicit — no spacing object needed.
+  - Non-default spacing MUST be an explicit spacing object with an ID.
+  - Zero spacing (overlap / flush) MUST be documented: ↕ `id` — space-zero (reason)
+  - Same spacing shared by all items in a group → define on the group, not between each item.
+  - Spacing objects flow into D-Design-System/00-design-system.md → Patterns.
+
+  FORMAT: #### ↕ `{id}` — {reason}
+
+  EXAMPLES:
+  #### ↕ `hem-v-space-zero` — header and service menu form one continuous nav unit
+  #### ↕ `hem-v-separator-2xl` — gray line, space-2xl above and below. Separates about from trust cards.
+  #### ↕ `hem-v-space-3xl` — major section boundary between seasons and footer
+-->
+
+---
+
+#### {Object Name 2}
+
+**OBJECT ID:** `{page-name}-{object-name-2}`
+
+| Property | Value |
+|----------|-------|
+| Component | [{Component}]({path}) |
+| Translation Key | `{translation.key}` |
+| SE | "{Swedish text}" |
+| EN | "{English text}" |
+
 ---
 
 #### {Group Name} (Container)
@@ -130,6 +214,8 @@
 | SE | "{Swedish text}" |
 | EN | "{English text}" |
 
+##### ↕ `{page-name}-{group-name}-{obj1}-{obj2}-gap` — {spacing token}
+
 ##### {Object in Group 2}
 
 **OBJECT ID:** `{page-name}-{group-name}-{object-2}`

package/src/workflows/4-ux-design/workflow-conceptualize.md

@@ -1,28 +1,23 @@
 ---
-name: 'workflow-conceptualize'
-description: 'Collaboratively explore what a scenario's design should achieve before committing to specific pages or layouts.'
+name: 'workflow-discuss'
+description: 'Creative dialog for page design — discuss what each page needs, then visualize and specify.'
 ---
 
-# [C] Conceptualize — Explore What the Design Needs
+# [C] Discuss — Creative Dialog for Page Design
 
-**Goal:** Collaboratively explore what a scenario's design should achieve before committing to specific pages or layouts.
+**Goal:** Lead a focused creative dialog for each page — what does it need, can we simplify it, then visualize and specify.
 
-**When to use:** When the user isn't sure what pages are needed, wants to think through the user journey, or needs to explore design options before building.
+**When to use:** The default design activity. Start here for any page that needs design thinking before building.
 
 ---
 
 ## INITIALIZATION
 
-### Agent Dialog Gate
-
-1. Check for pending activity dialogs
-2. If none, suggest creating one
-3. Load dialog context
-
+Read design log at `{output_folder}/_progress/00-design-log.md` before starting.
 
 ## Entry
 
-Load scenario context (VTC, scenario overview) from `{output_folder}/C-UX-Scenarios/`.
+Load scenario context (Trigger Map, scenario overview) from `{output_folder}/C-UX-Scenarios/`.
 
 ## Steps
 
@@ -33,6 +28,7 @@ Execute steps in `./steps-c/`:
 | 01 | step-01-exploration.md | Open-ended design exploration |
 
 **Reference data:**
+- `./data/guides/DESIGN-LOOP-GUIDE.md` — the 8-step design loop (discuss → wireframe → iterate → spec sync → implement → refine)
 - `./data/scenario-init/` — scenario initialization guides
 - `./data/page-creation-flows/` — page creation flow options
 
@@ -40,6 +36,4 @@ Execute steps in `./steps-c/`:
 
 ## AFTER COMPLETION
 
-1. Update design log
-2. Suggest next action
-3. Return to activity menu
+Step 01's two-option transitions handle all navigation. The design log is updated at every transition within the step itself. There is no separate "after completion" — the step loops through pages until the user stops or all pages are designed.

package/src/workflows/4-ux-design/workflow-design-system.md

@@ -13,16 +13,28 @@ description: 'Define, update, and review design system components used across pa
 
 ## INITIALIZATION
 
-### Agent Dialog Gate
+Read design log at `{output_folder}/_progress/00-design-log.md` before starting.
 
-1. Check for pending activity dialogs
-2. If none, suggest creating one
-3. Load dialog context
+## Extraction Rules
 
+Not everything extracts at the same time:
+
+### Objects: Extract on Second Use
+The first time a button, card, or widget appears, it stays inline in the page spec — it's a one-off. The **second** time the same pattern appears (same states, same behavior), it's a real pattern. Extract it to the design system.
+
+**First use = one-off. Second use = pattern. Extract.**
+
+### Spacing: Extract Immediately on First Use
+Spacing extracts on **first use** — no waiting for a second occurrence. Spacing is relational: when you decide that a heading needs `space-xl` above a card grid, that's a universal design principle, not a page-specific detail.
+
+### Component Extraction Check
+Before designing the 2nd+ page, scan completed specs for shared elements. If found, suggest extraction. Don't block the flow — the user can defer.
+
+---
 
 ## Entry
 
-Load design system from `{output_folder}/D-UX-Design/design-system/` (if exists).
+Load design system from `{output_folder}/D-Design-System/` (if exists).
 
 ## Steps
 
@@ -43,6 +55,6 @@ Execute steps in `./steps-m/`:
 
 ## AFTER COMPLETION
 
-1. Update design log
-2. Suggest next action
-3. Return to activity menu
+1. Append a progress entry to `{output_folder}/_progress/00-design-log.md` under `## Progress`:
+   `### [date] — Design System: [components extracted/updated]`
+2. Suggest next action based on the adaptive dashboard

package/src/workflows/4-ux-design/workflow-dream.md

@@ -13,41 +13,132 @@ description: 'The agent creates a complete scenario flow autonomously, then pres
 
 ## INITIALIZATION
 
-### Agent Dialog Gate
-
-1. Check for pending activity dialogs
-2. If none, suggest creating one
-3. Load dialog context
-
+Read design log at `{output_folder}/_progress/00-design-log.md` before starting.
 
 ## Entry
 
 Load scenario context from `{output_folder}/C-UX-Scenarios/`.
 
+### Scenario Check (CRITICAL GATE)
+
+Before starting page design, verify that a scenario exists for the selected scenario:
+
+1. Look for scenario files in `{output_folder}/C-UX-Scenarios/[NN-slug]/[NN-slug].md`
+2. **If a Phase 3 scenario exists** → Skip to **Process** below. The scenario's 8-question answers, shortest path, and first page specification provide everything needed.
+3. **If NO scenario exists** → Do NOT attempt to define the scenario here. Instead:
+   - Inform the user: *"Before we design pages, we need a scenario outline. This gives us the user's device, mental state, entry point, and the shortest path — all essential for good page design."*
+   - Suggest returning to Phase 3 to outline the scenario using the 8-question dialog
+   - The user can then return here with [D] from the Phase 3 post-scenario menu
+
+**Why:** Phase 3's 8-question dialog is the canonical way to define scenarios. It produces richer, more grounded scenarios than trying to shortcut the process here.
+
+### Phase 3 Handover Context
+
+When entering from Phase 3's [D] option (start designing), the scenario file and page folders already exist. Use:
+- **Page folders** from `{output_folder}/C-UX-Scenarios/[NN-slug]/pages/[NN].1-[page-slug]/` — each page has a boilerplate `.md` and a `Sketches/` subfolder
+- **First page spec** (`[NN].1-*.md`) has full entry context (device, arrival, mental state) from Q4, Q5, Q6
+- **Shortest path** from Q8 to know the full page sequence
+
 ## Process
 
 The Dream workflow uses the same steps as Suggest (`./steps-s/`) but with **autonomous execution**:
 
-1. **Agent executes all scenario setup steps** (step-01 through step-07) without pausing
-2. **Agent creates all pages** (step-08 through step-15) for each page in the flow
-3. **Agent presents the complete result** for user review
+1. **Agent creates all pages** (step-08 through step-15) for each page in the flow
+2. **Agent presents the complete result** for user review
 
 ### Agent Behavior
 
-- Make reasonable decisions at each step based on VTC, scenario context, and WDS patterns
+- Make reasonable decisions at each step based on Trigger Map, scenario context, and WDS patterns
 - Document decisions and rationale as you go
 - When uncertain, choose the simpler option
 - After completion, present a summary of all decisions made
 - User can accept, request changes, or switch to **[S] Suggest** for finer control
 
+### Mode Override Rule (CRITICAL)
+
+Step files in `./steps-s/` contain rules like "ALWAYS halt and wait for user input" and "NEVER generate content without user input." **These rules apply ONLY in Suggest mode.**
+
+In Dream mode:
+- **OVERRIDE** all "halt and wait" rules — auto-proceed after completing each step
+- **OVERRIDE** "NEVER generate content without user input" — generate based on context and WDS patterns
+- **DO NOT** display menus or wait for menu selections between steps
+- **DO** still save outputs and update the design log at each step
+- **DO** still follow the step's actual instructions for what to generate
+- The user can type **"stop"** or **"pause"** at any time to interrupt and switch to Suggest mode
+
 **Reference data:**
 - `./data/scenario-init/` — scenario guides and examples
 - `./data/page-creation-flows/` — page creation approaches
 
 ---
 
+## DESIGN LOG REPORTING
+
+In Dream mode, the agent updates the design log autonomously at each page completion. Append to the Design Loop Status table in `{output_folder}/_progress/00-design-log.md`:
+
+```
+| [Scenario slug] | [NN.X] | [Page name] | specified | [YYYY-MM-DD] |
+```
+
+Do NOT skip this — even in autonomous mode, the design log must be updated per page.
+
 ## AFTER COMPLETION
 
-1. Update design log
-2. Suggest next action
-3. Return to activity menu
+### Autonomous Mode (all pages at once)
+
+When Dream mode completes all pages in the scenario, present a summary for review:
+
+<output>
+**Dream complete! Here's what I created for [Scenario Name]:**
+
+| Step | Page | Status | Key Decisions |
+|------|------|--------|---------------|
+| [NN.1] | [page name] | specified | [brief summary] |
+| [NN.2] | [page name] | specified | [brief summary] |
+| ... | ... | ... | ... |
+
+**Shared components extracted:** [list if any]
+
+Review the pages and let me know what to adjust. When you're happy:
+
+1. **Start building** — hand the first page to agentic development
+2. **Explore responsive states / storyboard** — if any pages need detail work
+</output>
+
+### Per-Page Mode (user interrupted or reviewing one at a time)
+
+Present the same two-option transition as Discuss and Suggest:
+
+**If complexity exists on this page:**
+
+<output>
+**Specification for "[page name]" is complete.**
+
+This page has [responsive states / storyboard items / complex functionality] that need exploring.
+
+1. **Explore [responsive states / storyboard / functionality]** — define the details
+2. **Explore the next scenario step** — [next page name]
+</output>
+
+**If simple page:**
+
+<output>
+**Specification for "[page name]" is complete. Ready to build.**
+
+1. **Build it** — start agentic development
+2. **Explore the next scenario step** — [next page name]
+</output>
+
+### Component Extraction (Dream Mode)
+
+In Dream mode, component extraction runs automatically:
+1. Scan completed page specs silently after each page
+2. If shared elements found, auto-extract as shared components (log decisions)
+3. Reference shared components in subsequent page specs instead of duplicating
+4. Include extraction summary in the final review presentation
+
+### Execution Rules
+
+- ALWAYS halt and wait for user input after presenting review/transition
+- User can type "stop" or "pause" to interrupt autonomous mode
+- The user can always say "stop" to pause and return later — log current status

package/src/workflows/4-ux-design/workflow-handover.md

@@ -13,11 +13,7 @@ description: Package complete testable flows and hand off to development
 
 ## INITIALIZATION
 
-### Agent Dialog Gate
-
-1. Check `{output_folder}/_progress/agent-dialogs/` for pending handover dialogs
-2. If none exists, suggest creating one for this handover
-3. If exists, load and resume
+Read design log at `{output_folder}/_progress/00-design-log.md` before starting.
 
 ---
 
@@ -43,6 +39,6 @@ Execute steps in `./steps-h/`:
 
 ## AFTER COMPLETION
 
-1. Update design log
-2. Suggest next agent dialog action
-3. Return to scenario dashboard
+1. Append a progress entry to `{output_folder}/_progress/00-design-log.md` under `## Progress`:
+   `### [date] — Design Delivery: [what was packaged]`
+2. Suggest next action: Phase 5 prototyping or next scenario

package/src/workflows/4-ux-design/workflow-sketch.md

@@ -13,12 +13,7 @@ description: 'Analyze user-provided sketches (photos, screenshots, wireframes) a
 
 ## INITIALIZATION
 
-### Agent Dialog Gate
-
-1. Check for pending activity dialogs
-2. If none, suggest creating one
-3. Load dialog context
-
+Read design log at `{output_folder}/_progress/00-design-log.md` before starting.
 
 ## Entry
 
@@ -41,6 +36,4 @@ Execute steps in `./steps-k/`:
 
 ## AFTER COMPLETION
 
-1. Update design log
-2. Suggest next action
-3. Return to activity menu
+After sketch analysis, the page returns to step-01-exploration.md's flow. The sketch analysis feeds into the wireframe/spec sync step — the calling step handles design log updates and transition options.