zhongwei/gh-webdevtodayjason-titanium-plugins-plugins-titanium-toolkit

Files

Zhongwei Li db12a906d2 Initial commit

2025-11-30 09:05:52 +08:00

25 KiB

Raw Permalink Blame History

description

description
Complete guided BMAD backlog generation workflow

BMAD Start - Complete Guided Workflow

You are orchestrating complete BMAD backlog generation from idea to implementation-ready documentation. This guided workflow creates Product Brief, PRD, Architecture, all Epic files, and Story Index in one comprehensive session.

MCP Tools Used: This command uses the tt MCP server (Titanium Toolkit) which provides:

mcp__plugin_titanium-toolkit_tt__bmad_generator - Generates BMAD documents (brief, PRD, architecture, epics, index)
mcp__plugin_titanium-toolkit_tt__bmad_validator - Validates BMAD document structure and completeness

The tt server wraps Python utilities that use Claude AI to generate comprehensive project documentation following the BMAD methodology.

Purpose

Take user from empty folder to complete project backlog (30-45 minutes):

Product Brief (high-level vision)
PRD (comprehensive requirements with epics and user stories)
Architecture (technical design with code examples)
Epic files (all epics with detailed stories and acceptance criteria)
Story Index (summary of all stories)

Overview

This workflow has 6 phases:

Introduction - Welcome and mode selection
Product Brief - Capture high-level vision
PRD - Comprehensive requirements
Research - Technical research (if needed)
Architecture - Technical design
Epics & Index - All epic files and summary

Duration: 30-45 minutes Cost: ~$0.15 (Claude Haiku 4.5 + Sonnet 4.5 for all documents) Output: Complete bmad-backlog/ folder

Phase 1: Introduction & Setup

Step 1.1: Welcome User

🚀 BMAD Document Generation - Complete Guided Workflow

Welcome to the BMAD backlog creation process!

This workflow will create complete project documentation:
✅ Product Brief - High-level vision and goals
✅ PRD - Comprehensive requirements with epics
✅ Architecture - Technical design with code examples
✅ Epic Files - Detailed user stories with acceptance criteria
✅ Story Index - Summary of all stories for sprint planning

⏱️  Duration: 30-45 minutes
💰 Cost: ~$0.15 (Claude for document generation)
📁 Output: bmad-backlog/ folder with all documentation

Ready to start? (yes/no)

If no → Exit If yes → Continue

Step 1.2: Check for Existing Documentation

ls -la bmad-backlog/ 2>/dev/null || echo "No backlog found"

If bmad-backlog/ exists:

⚠️  Found existing BMAD documentation!

bmad-backlog/ folder exists with:
{{List existing files}}

Options:
1. **Start fresh** - Delete existing and create new (DESTRUCTIVE)
2. **Add to existing** - Generate new epics only
3. **Cancel** - Review existing docs first

What would you like to do?

If option 1 chosen:

🛑 DESTRUCTIVE ACTION WARNING

This will permanently DELETE the entire bmad-backlog/ folder including:
- Product Brief
- PRD
- Architecture
- All Epic files
- Story Index
- Research findings

This CANNOT be undone.

To confirm deletion, type exactly: DELETE

Type DELETE to confirm, or anything else to cancel:

Wait for user input.

If user types exactly "DELETE" (case-sensitive):

rm -rf bmad-backlog/

Announce: "bmad-backlog/ deleted. Starting fresh..."
Continue to Step 1.3

If user types anything else:

❌ Deletion cancelled. Existing backlog preserved.

Returning to options menu...

Go back to showing the 3 options
Wait for user to choose option 2 or 3

If option 2: Skip to Phase 6 (epic generation) If option 3: Exit gracefully

If no bmad-backlog/:

Continue to Step 1.3

Step 1.3: Choose Mode

Choose your workflow mode:

1. **Interactive** (Recommended)
   - I'll ask questions section by section
   - You review and refine as we go
   - Higher quality, more control
   - ~45 minutes

2. **YOLO** (Faster)
   - I'll generate complete drafts from your idea
   - You refine afterwards
   - Faster but may need more editing
   - ~30 minutes

Which mode? (1 or 2)

Store user's choice for the workflow.

Phase 2: Product Brief Generation

Step 2.1: Capture High-Level Idea

🎯 Let's start with your project idea.

Describe your project at a high level:
- What is it?
- Who is it for?
- What problem does it solve?

You can be brief (one sentence) or detailed (a paragraph).
I'll ask follow-up questions to flesh it out.

Your idea:

Wait for user response. Store as user_idea.

Step 2.2: Interactive Questions (if Interactive mode)

Ask clarifying questions:

"What problem does this solve for users?"
"Who are the primary users (be specific)?"
"What makes this different from existing solutions?"
"What are the 3-5 core features for MVP?"
"What should be out of scope for MVP?"
"Any technical preferences (language, framework, hosting)?"
"Budget and timeline constraints?"

Append answers to user_idea context.

Step 2.3: Launch Product-Manager Subagent for Brief

Use the Task tool to launch the product-manager subagent:

Task(
  description: "Generate BMAD product brief",
  prompt: "Create comprehensive product brief following BMAD methodology.

User's Project Idea:
{{user_idea}}

Your workflow:

1. **Generate product brief** using MCP tool:

mcp__plugin_titanium-toolkit_tt__bmad_generator( doc_type: "brief", input_path: "{{user_idea}}", project_path: "$(pwd)" )


2. **Review and present** key sections to user:
- Executive Summary
- Problem Statement
- MVP Scope
- Ask for approval or refinements

3. **Validate the brief** using:

mcp__plugin_titanium-toolkit_tt__bmad_validator( doc_type: "brief", document_path: "bmad-backlog/product-brief.md" )


4. **Run vibe-check** to validate brief quality

5. **Store in Pieces** for future reference

6. **Present summary** confirming brief is complete

**IMPORTANT**: Keep your summary response BRIEF (under 300 tokens). The orchestrator just needs:
- Confirmation brief is complete
- 1-2 sentence project description
- Primary user segment
- MVP feature count

DO NOT include the full brief content in your response - it's already saved to the file.

Follow your complete brief workflow from the bmad-methodology skill.

Project path: $(pwd)",
subagent_type: "product-manager"
)

The product-manager subagent will handle brief generation, review, validation, and storage.

Step 2.4: Confirm Completion

Wait for product-manager to return summary. Confirm brief is ready before continuing to PRD.

Phase 3: PRD Generation

Step 3.1: Launch Product-Manager Subagent for PRD

Generating comprehensive Product Requirements Document...

This will take about 1-2 minutes (large document).

Use the Task tool to launch the product-manager subagent:

Task(
  description: "Generate BMAD PRD",
  prompt: "Create comprehensive Product Requirements Document following BMAD methodology.

Input:
- Product Brief: bmad-backlog/product-brief.md

Output:
- PRD: bmad-backlog/prd/prd.md

Your workflow:

1. **Read product brief** to understand context

2. **Generate PRD** using MCP tool:

mcp__plugin_titanium-toolkit_tt__bmad_generator( doc_type: "prd", input_path: "bmad-backlog/product-brief.md", project_path: "$(pwd)" )


3. **Review epic structure**:
- Read generated PRD
- Extract epic list
- Count epics, user stories, features
- Verify Epic 1 is Foundation/Infrastructure
- Present structure to user for approval

4. **Validate PRD** using:

mcp__plugin_titanium-toolkit_tt__bmad_validator( doc_type: "prd", document_path: "bmad-backlog/prd/prd.md" )


5. **Run vibe-check** with:
- Goal: Create comprehensive PRD
- Plan: Generated PRD with N epics, M stories
- Uncertainties: Epic structure, requirement completeness, missing features

6. **Store in Pieces** for future reference

7. **Present summary** with epic list and totals

**IMPORTANT**: Keep your summary response BRIEF (under 500 tokens). The orchestrator just needs:
- Confirmation PRD is complete
- Epic count and list (just titles)
- Total user stories count
- Total features count

DO NOT include the full PRD content in your response - it's already saved to the file.

Follow your complete PRD workflow from the bmad-methodology skill.

Project path: $(pwd)",
subagent_type: "product-manager"
)

The product-manager subagent will handle PRD generation, epic structure review, validation, and storage.

Step 3.2: Confirm Completion

Wait for product-manager to return summary. Confirm PRD is ready and epic structure approved before continuing to research/architecture.

Phase 4: Research (If Needed)

Step 4.1: Detect Research Needs

Scan PRD for keywords:

"API", "vendor", "data source"
"payment", "authentication"
"hosting", "infrastructure"

If research topics detected:

🔍 Research Opportunities Detected!

Your PRD mentions:
- {{Topic 1}} (e.g., "data vendors")
- {{Topic 2}} (e.g., "authentication provider")
- {{Topic 3}} (e.g., "hosting platform")

These decisions will impact your architecture.

Would you like research prompts for these? (yes/no/specific topics)

Step 4.2: Generate Research Prompts

For each selected topic: Run /bmad:research "{{topic}}" for each

Present:

📋 Research Prompts Generated!

I've created research prompts for:
1. {{Topic 1}} - bmad-backlog/research/RESEARCH-{{slug}}-prompt.md
2. {{Topic 2}} - bmad-backlog/research/RESEARCH-{{slug}}-prompt.md

Please:
1. Open each prompt file
2. Copy to ChatGPT/Claude web
3. Complete research (~10-15 min per topic)
4. Fill in findings templates
5. Return here when done

Ready to continue after research? (yes/skip)

If yes: Wait for confirmation research is complete If skip: Continue without research

Phase 5: Architecture Generation

Step 5.1: Launch Architect Subagent

Generating Technical Architecture...

This will take 2-3 minutes (comprehensive document with code examples).

Use the Task tool to launch the architect subagent:

Task(
  description: "Generate BMAD architecture",
  prompt: "Create comprehensive technical architecture document following BMAD methodology.

Input:
- PRD: bmad-backlog/prd/prd.md
- Research findings: bmad-backlog/research/*.md (if any exist)

Output:
- Architecture document: bmad-backlog/architecture/architecture.md

Your workflow:

1. **Read PRD** to understand requirements

2. **Check for research findings** and incorporate recommendations if they exist

3. **Generate architecture** using MCP tool:

mcp__plugin_titanium-toolkit_tt__bmad_generator( doc_type: "architecture", input_path: "bmad-backlog/prd/prd.md", project_path: "$(pwd)" )


4. **Review tech stack with user**:
- Present proposed tech stack
- Note research-based decisions if applicable
- Get user approval or note requested changes

5. **Validate architecture** using:

mcp__plugin_titanium-toolkit_tt__bmad_validator( doc_type: "architecture", document_path: "bmad-backlog/architecture/architecture.md" )


6. **Run vibe-check** to validate architectural decisions

7. **Store in Pieces** for future reference

8. **Present summary** with tech stack and cost estimates

**IMPORTANT**: Keep your summary response BRIEF (under 500 tokens). The orchestrator just needs:
- Confirmation architecture is complete
- Proposed tech stack (2-3 sentences)
- MVP cost estimate
- Any critical decisions made

DO NOT include the full architecture content in your response - it's already saved to the file.

Follow your complete architecture workflow from the bmad-methodology skill.

Project path: $(pwd)",
subagent_type: "architect"
)

The architect subagent will handle architecture generation, tech stack review, validation, and storage.

Step 5.2: Confirm Completion

Wait for architect to return summary. Confirm architecture is ready before continuing to epic generation.

Phase 6: Epic Generation & Index

Step 6.1: Extract Epic List from PRD

Read PRD and identify all epics from User Stories section.

Count epics:

Found {{N}} epics in PRD:
1. {{Epic name}}
2. {{Epic name}}
3. {{Epic name}}
...

Generating all {{N}} epic files...

Step 6.2: Generate All Epics (Delegated to Product-Manager)

For each epic, launch product-manager subagent (sequential):

Task(
  description: "Generate BMAD epic {{epic_num}}: {{epic_name}}",
  prompt: "Create comprehensive epic file following BMAD methodology.

Epic to Generate: {{epic_num}}

Input:
- PRD: bmad-backlog/prd/prd.md
- Architecture: bmad-backlog/architecture/architecture.md

Output:
- Epic file: bmad-backlog/epics/EPIC-{num:03d}-{slug}.md

Your workflow:

1. **Read inputs** to understand context

2. **Generate epic** using MCP tool:

mcp__plugin_titanium-toolkit_tt__bmad_generator( doc_type: "epic", input_path: "bmad-backlog/prd/prd.md bmad-backlog/architecture/architecture.md {{epic_num}}", project_path: "$(pwd)" )


3. **Present brief summary**:
- Epic title
- Story count
- Story points

4. **Validate epic** using:

mcp__plugin_titanium-toolkit_tt__bmad_validator( doc_type: "epic", document_path: "bmad-backlog/epics/EPIC-{num}-{name}.md" )


5. **Run vibe-check** to validate epic quality

6. **Store in Pieces** for future reference

7. **Return summary** (brief - we're in batch mode)

**IMPORTANT**: Keep your summary response VERY BRIEF (under 200 tokens). Just return:
- "Epic {{num}} complete: {{title}} ({{story_count}} stories)"

DO NOT include full epic content - it's already saved. We're generating multiple epics so be concise.

Follow your complete epic workflow from the bmad-methodology skill.

Project path: $(pwd)",
subagent_type: "product-manager"
)

Show progress after each:

Generating epics...
✅ Epic 1: {{name}} ({{X}} stories)
✅ Epic 2: {{name}} ({{Y}} stories)
✅ Epic 3: {{name}} ({{Z}} stories)
⏳ Epic 4: {{name}} (generating...)

Step 6.3: Generate Story Index (Delegated to Product-Manager)

After all epics, launch product-manager subagent for index:

Task(
  description: "Generate BMAD story index",
  prompt: "Create comprehensive story index summarizing all epics and user stories.

Input:
- Epic files: bmad-backlog/epics/EPIC-*.md

Output:
- Story index: bmad-backlog/STORY-INDEX.md

Your workflow:

1. **Generate story index** using MCP tool:

mcp__plugin_titanium-toolkit_tt__bmad_generator( doc_type: "index", input_path: "bmad-backlog/epics/", project_path: "$(pwd)" )


2. **Extract totals**:
- Total epics, stories, story points
- Epic breakdown
- Priority distribution

3. **Run vibe-check** to validate index quality

4. **Store in Pieces** for future reference

5. **Return summary** with totals

Follow your complete index workflow from the bmad-methodology skill.

Project path: $(pwd)",
subagent_type: "product-manager"
)

Creates: bmad-backlog/STORY-INDEX.md

Phase 7: Final Review & Summary

Step 7.1: Confirm All Documents Created

Verify all expected documents exist:

# Check core documents
ls bmad-backlog/product-brief.md
ls bmad-backlog/prd/prd.md
ls bmad-backlog/architecture/architecture.md
ls bmad-backlog/STORY-INDEX.md

# Count epic files
ls bmad-backlog/epics/EPIC-*.md | wc -l

Note: Each subagent already validated their documents during generation. This is just a final check that all files were created.

Step 7.2: Final vibe-check

mcp__vibe-check__vibe_check(
  goal: "Create complete BMAD backlog for {{project}}",
  plan: "Generated: Product Brief, PRD ({{N}} epics, {{M}} stories), Architecture ({{tech stack}}), {{N}} Epic files, Story Index",
  uncertainties: [
    "Is documentation complete and consistent?",
    "Are epics logically sequenced?",
    "Is architecture aligned with requirements?",
    "Any gaps in the backlog?"
  ]
)

Present vibe-check assessment.

Step 7.3: Store Complete Backlog in Pieces

First, build explicit file list (Pieces doesn't expand wildcards):

# Use Glob to find all epic files
Glob bmad-backlog/epics/EPIC-*.md

Store the results, then build files array.

If research was completed:

# Find research files
Glob bmad-backlog/research/RESEARCH-*.md

Then call Pieces with explicit paths:

mcp__Pieces__create_pieces_memory(
  summary_description: "Complete BMAD backlog for {{project}}",
  summary: "Generated complete project backlog using BMAD methodology.

Documents created:
- Product Brief: {{vision summary}}
- PRD: {{N}} sections, {{E}} epics, {{M}} user stories, {{F}} features
- Architecture: Tech stack ({{stack}}), database schemas ({{table count}} tables), infrastructure plan ({{hosting}}), security architecture
- {{E}} Epic files: Each with stories, acceptance criteria, technical notes
- Story Index: {{M}} stories, {{P}} story points

Key features:
{{List main features}}

Tech stack:
{{Stack summary}}

Epics:
{{List all epics}}

Cost estimates:
- MVP: ${{X}}/month
- Production: ${{Y}}/month

Ready for implementation with /titanium:plan bmad-backlog/epics/EPIC-001-*.md",
  files: [
    "bmad-backlog/product-brief.md",
    "bmad-backlog/prd/prd.md",
    "bmad-backlog/architecture/architecture.md",
    "{{list each epic file from Glob results}}",
    "bmad-backlog/STORY-INDEX.md",
    "{{list each research file from Glob results if research completed}}"
  ],
  project: "$(pwd)"
)

Example with explicit paths:

files: [
  "bmad-backlog/product-brief.md",
  "bmad-backlog/prd/prd.md",
  "bmad-backlog/architecture/architecture.md",
  "bmad-backlog/epics/EPIC-001-foundation.md",
  "bmad-backlog/epics/EPIC-002-core-features.md",
  "bmad-backlog/epics/EPIC-003-advanced.md",
  "bmad-backlog/epics/EPIC-004-polish.md",
  "bmad-backlog/STORY-INDEX.md"
]

If research completed, add:

  "bmad-backlog/research/RESEARCH-data-vendors-prompt.md",
  "bmad-backlog/research/RESEARCH-data-vendors-findings.md"

Step 7.4: Present Complete Summary

🎉 BMAD Backlog Generation Complete!

📚 Documentation Created:

✅ Product Brief
   Location: bmad-backlog/product-brief.md
   Purpose: High-level vision and goals

✅ Product Requirements Document (PRD)
   Location: bmad-backlog/prd/prd.md
   Size: ~{{lines}} lines
   Epics: {{N}}
   User Stories: {{M}}
   Features: {{F}}

✅ Technical Architecture
   Location: bmad-backlog/architecture/architecture.md
   Size: ~{{lines}} lines
   Tech Stack: {{stack summary}}
   Database Tables: {{count}}
   Cost Estimate: ${{MVP cost}}/month MVP

✅ Epic Files ({{N}} epics)
   Location: bmad-backlog/epics/
   Files:
   - EPIC-001-{{name}}.md ({{X}} stories)
   - EPIC-002-{{name}}.md ({{Y}} stories)
   - EPIC-003-{{name}}.md ({{Z}} stories)
   ...

✅ Story Index
   Location: bmad-backlog/STORY-INDEX.md
   Total: {{M}} stories, {{P}} story points

{{If research completed:}}
✅ Research Findings
   Location: bmad-backlog/research/
   Topics: {{list topics}}

---

📊 Project Summary:

Goal: {{high-level goal}}
Target Users: {{primary user segment}}
MVP Features: {{core feature count}}
Tech Stack: {{Frontend}} + {{Backend}} + {{Database}}
MVP Timeline: 12-16 weeks (from PRD estimate)
Cost: ${{X}}/month MVP, ${{Y}}/month Production

Epic Sequence:
1. {{Epic 1}} - Foundation ({{stories}} stories, {{points}} points)
2. {{Epic 2}} - {{name}} ({{stories}} stories, {{points}} points)
3. {{Epic 3}} - {{name}} ({{stories}} stories, {{points}} points)
...

Total Development Effort: {{total story points}} story points

---

📁 Folder Structure:

bmad-backlog/
├── product-brief.md
├── prd/
│   └── prd.md
├── architecture/
│   └── architecture.md
├── epics/
│   ├── EPIC-001-{{name}}.md
│   ├── EPIC-002-{{name}}.md
│   └── ... ({{N}} total)
{{If research:}}
├── research/
│   ├── RESEARCH-{{topic}}-prompt.md
│   └── RESEARCH-{{topic}}-findings.md
└── STORY-INDEX.md

---

💡 Next Steps:

Ready to start implementing?

**Option 1: Plan First Epic** (Recommended)

/titanium:plan bmad-backlog/epics/EPIC-001-{{name}}.md

This will break Epic 1 into implementation tasks.

**Option 2: Review Documentation First**
Open and review:
- bmad-backlog/prd/prd.md (requirements)
- bmad-backlog/architecture/architecture.md (technical design)
- bmad-backlog/epics/*.md (user stories)

Then come back and run:

/titanium:plan bmad-backlog/epics/EPIC-001-{{name}}.md


**Option 3: Complete Workflow**

/titanium:work bmad-backlog/epics/EPIC-001-{{name}}.md

This will plan AND implement Epic 1 in one go.

---

🎊 Congratulations! You now have complete project documentation following BMAD methodology.

Your backlog is ready for professional AI-powered development with Titanium Toolkit.

What would you like to do next?

Detailed Workflow

Interactive Mode Steps

Product Brief:

Capture idea
Ask 7-10 clarifying questions
Generate brief
Show to user
Refine if needed
Validate and approve

PRD:

Read brief
Ask about success metrics
Ask about features (work through MVP features)
Propose epic structure
For each epic, ask about user stories
Ask about technical requirements
Ask about data/AI needs
Ask about design preferences
Ask about go-to-market
Ask about risks
Generate comprehensive PRD
Review epic list with user
Validate and approve

Research (if needed):

Detect research needs from PRD
Offer to generate prompts
Generate prompts and templates
Wait for user to research
Confirm research complete
Continue

Architecture:

Read PRD
Read research findings (if exist)
Propose tech stack
Get user approval
Generate comprehensive architecture
Show tech stack and cost estimates
Validate and approve

Epics:

Extract epic list from PRD
For each epic:
- Generate epic file
- Show story count
- Progress indicator
Generate story index
Show totals

YOLO Mode Steps

All at once:

Capture idea
Ask 2-3 critical questions only
Generate ALL documents with one set of prompts
Present complete backlog
User reviews and refines any sections needed

Faster but less precise - good for experienced users.

Error Handling

If ANTHROPIC_API_KEY Missing

❌ Error: ANTHROPIC_API_KEY required

This workflow uses Claude (Haiku 4.5 + Sonnet 4.5) to generate comprehensive documentation.

Add your key to ~/.env:
  echo 'ANTHROPIC_API_KEY=sk-ant-your-key' >> ~/.env
  chmod 600 ~/.env

Restart Claude Code and try again.

Cost: ~$0.15 for complete backlog (less than a coffee!)

If Generation Fails Mid-Workflow

❌ Error during {{document}} generation

Error: {{error message}}

Options:
1. Retry this step
2. Skip this document (not recommended)
3. Cancel workflow (documents created so far are saved)

What would you like to do?

If User Cancels Mid-Workflow

⏸️  Workflow Paused

Documents created so far:
{{List what's been created}}

You can resume by running individual commands:
- /bmad:prd (if brief complete)
- /bmad:architecture (if PRD complete)
- /bmad:epic {{N}} (if architecture complete)

Or start over: /bmad:start

Voice Feedback

Voice announces progress:

"Starting BMAD workflow" (at beginning)
"Product brief complete" (after phase 2)
"PRD complete: {{N}} epics" (after phase 3)
"Architecture complete" (after phase 5)
"Generating epics" (phase 6)
"Epic {{N}} of {{total}} complete" (each epic)
"Backlog complete: {{M}} stories" (at end)

Time Estimates

Interactive Mode:

Product Brief: ~10 minutes (questions + generation)
PRD: ~15 minutes (questions + generation)
Research: ~15 minutes per topic (optional)
Architecture: ~5 minutes (generation + review)
Epics: ~2 minutes per epic (generation)
Total: ~45 minutes (without research) to ~60 minutes (with research)

YOLO Mode:

Idea capture: ~5 minutes
Generation: ~10 minutes (all documents)
Review: ~15 minutes
Total: ~30 minutes

Important Guidelines

Always:

✅ Validate each document before proceeding
✅ Use vibe-check for quality validation
✅ Store progress in Pieces after each phase
✅ Present clear summaries
✅ Verify Epic 1 is Foundation
✅ Offer research opportunities
✅ Guide user to next steps

Never:

❌ Skip document validation
❌ Ignore vibe-check concerns
❌ Generate incomplete documents
❌ Forget to generate story index
❌ Leave user uncertain about next steps

Integration

After /bmad:start completes:

/titanium:plan bmad-backlog/epics/EPIC-001-*.md
/titanium:work

Complete cycle:

Idea → /bmad:start → /titanium:plan → /titanium:work → Working Code
  (45 min)        (2 min)         (3-6 hours)

Cost Breakdown:

BMAD generation: $0.22
Implementation: $0.10 per epic
Total for MVP: $0.22 + (4 epics × $0.10) = ~$0.62

This command transforms an idea into a complete, implementation-ready project backlog in under an hour!

25 KiB Raw Permalink Blame History Unescape Escape