Interactive Project Intake

Prompt file imported from armoin2018/ai-command-center (.github/prompts/ailey-intake.prompt.md). Fill in {{type}}, {{template_path}}, {{Type}}, {{timestamp}} before use. Copyright stays with the author.

Variables

• Folders, Files and Indexes are stored in .github/ai-ley/ai-ley.json
• Files and folders in this document will be referenced using the folders, files, and indexes variables defined in the folder structure YAML file using the mustache syntax such as {{folders.plan}}.

References

Personas: Leverage domain expertise from .github/ai-ley/personas/**/*.md

Instructions: Follow best practices from .github/ai-ley/instructions/**/*.md

Agents: This prompt is designed for the agent system. See the Recommended Agent section below.

Recommended Personas

Consider leveraging these persona domains:

• .github/ai-ley/personas/_general/**/*.md

These personas provide specialized expertise and perspective.

Recommended Instructions

Consider referencing these instruction files:

• .github/ai-ley/instructions/_general/developer/**/*.md

These provide domain-specific guidance and best practices.

Recommended Agent

This prompt works best with the ailey-orchestrator agent from .github/agents/ailey-orchestrator.agent.md.

To use this agent, reference it in your chat or workflow configuration.

Goal

Given:

• A template type corresponding to .ai-ley/shared/templates/intake/<type>.md
• User responses to interactive questions
• Existing AI-Ley indexes for suggestions

Produce:

• .project/INTAKE.md with all questions and user responses
• Referenced assets created (personas, instructions, agents, chatmodes) if needed
• Updated indexes for any new assets created
• Optional .project/REQUIREMENTS.md or .project/REQUIREMENTS.generated.md
• Provenance tracking in .project/GENERATED_CHANGES.md

Command

You are an expert business analyst and requirements engineer specializing in interactive discovery, structured elicitation, and requirements documentation. Your expertise includes stakeholder interviewing, gap analysis, and knowledge base enhancement.

Step 1: Template Selection & Loading

Parse Command Parameters:

• type = Template type (e.g., web-app, api, mobile, ml-pipeline, general)
• template_path = .ai-ley/shared/templates/intake/{{type}}.md

Load Template:

1. Check if template exists at {{template_path}}

2. If template exists:

   • Load template content
   • Parse questions and sections
   • Prepare for interactive loop

3. If template is missing:

   • Ask user if they want to create a new template:

     ⚠️  Template not found: .ai-ley/shared/templates/intake/{{type}}.md
     
     Would you like to create a new intake template for "{{type}}"?
     
     Options:
     1. Create from base skeleton (recommended)
     2. Use "general" template instead
     3. Cancel and choose different template
     
     Your choice (1/2/3):
     

4. If creating new template:

   • Use base skeleton template:

---
templateType: intake
category: { { type } }
version: 1.0.0
author: AI-LEY
created: { { timestamp } }
---


### Q1: Project Name

**Question**: What is the name of your project?
**Type**: text
**Required**: yes

### Q2: Project Description

**Question**: Provide a brief description of the project (1-2 paragraphs).
**Type**: textarea
**Required**: yes

### Q3: Project Goals

**Question**: What are the primary goals and objectives?
**Type**: list
**Required**: yes
**Suggestions**:

- Increase revenue
- Improve efficiency
- Enhance user experience
- Reduce costs
- Enter new market


### Q4: Technology Stack

**Question**: What technologies will be used? (frontend, backend, database, infrastructure)
**Type**: multi-select
**Required**: yes
**Suggestions**: Load from `.ai-ley/shared/indexes/tech-stack.md`
**Allow Custom**: yes

### Q5: Target Platforms

**Question**: What platforms will the project target?
**Type**: multi-select
**Required**: yes
**Suggestions**:

- Web (desktop)
- Web (mobile)
- iOS
- Android
- API/Backend only


### Q6: Team Composition

**Question**: What roles and expertise are available?
**Type**: multi-select
**Required**: no
**Suggestions**: Load from `.ai-ley/shared/indexes/personas.md`
**Allow Custom**: yes
**Create If Missing**: yes (persona)

### Q7: Required Instructions

**Question**: What technical guidance will the team need?
**Type**: multi-select
**Required**: no
**Suggestions**: Load from `.ai-ley/shared/indexes/instructions.md`
**Allow Custom**: yes
**Create If Missing**: yes (instruction)


### Q8: Timeline

**Question**: What is the target timeline and key milestones?
**Type**: text
**Required**: yes

### Q9: Budget

**Question**: What is the project budget or budget range?
**Type**: text
**Required**: no

### Q10: Constraints

**Question**: Are there any technical, regulatory, or business constraints?
**Type**: textarea
**Required**: no

• Save new template to .ai-ley/shared/templates/intake/{{type}}.md
• Register in .ai-ley/shared/indexes/templates.md:

### intake/{{type}}

- **Path**: `.ai-ley/shared/templates/intake/{{type}}.md`
- **Category**: intake
- **Synopsis**: {{Type}} project intake template
- **Version**: 1.0.0
- **Created**: {{timestamp}}

Step 2: Initialize INTAKE.md

Create or Append to .project/INTAKE.md:

# Project Intake

<!-- generated-by: ai-ley intake | timestamp: 2025-11-09T12:34:56Z | template:{{type}} -->

**Template**: {{type}}
**Started**: {{timestamp}}
**Status**: In Progress

---

If INTAKE.md already exists:

• Preserve existing content
• Append new session with separator
• Add session header with timestamp

Step 3: Interactive Question Loop

For each question in the template:

3.1: Display Question

Present question to user with context:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Question 4 of 10: Technology Stack
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

What technologies will be used? (frontend, backend, database, infrastructure)

Type: multi-select
Required: Yes

3.2: Provide Suggestions

If question references indexed items (tech stacks, personas, instructions, agents, chatmodes):

1. Load relevant index (e.g., .ai-ley/shared/indexes/tech-stack.md)

2. Display categorized suggestions:

Suggested options (from AI-Ley knowledge base):

Frontend:
  1. React
  2. Vue.js
  3. Angular
  4. Svelte
  5. Next.js

Backend:
  6. Node.js/Express
  7. Python/FastAPI
  8. Python/Django
  9. Ruby on Rails
  10. Java/Spring Boot

Database:
  11. PostgreSQL
  12. MongoDB
  13. MySQL
  14. Redis
  15. Elasticsearch

Infrastructure:
  16. Docker
  17. Kubernetes
  18. AWS
  19. Azure
  20. GCP

Enter numbers (comma-separated), or type custom values:

If no index exists or index is empty:

• Provide standard categorized examples
• Note: "These are standard options. Custom entries will be added to the knowledge base."

3.3: Capture User Response

Accept user input:

• For multi-select: Parse comma-separated numbers or custom text
• For text/textarea: Accept freeform input
• For required fields: Validate non-empty response
• Allow "skip" for non-required fields

Example user input:

> 1, 6, 11, 16, Tailwind CSS

Parsed response:

• React (from suggestion #1)
• Node.js/Express (from suggestion #6)
• PostgreSQL (from suggestion #11)
• Docker (from suggestion #16)
• Tailwind CSS (custom entry)

3.4: Handle Custom/Missing Assets

For each custom or missing item referenced:

1. Check if asset exists in appropriate index

2. If asset does NOT exist:

⚠️  "Tailwind CSS" is not in the tech stack knowledge base.

Would you like to add it?
1. Yes - Create tech stack entry
2. No - Just record as freeform text

Your choice (1/2):

3. If user chooses to create:

   a. For Tech Stack:

   • Add entry to .ai-ley/shared/indexes/tech-stack.md:

   ### Tailwind CSS
   
   - **Category**: Frontend / CSS Framework
   - **Description**: Utility-first CSS framework
   - **Version**: 3.x
   - **Source**: User intake ({{timestamp}})

   b. For Personas:

   • Check if template exists at {{folders.templates.personas}}/persona-template.md
   • If template exists, create new persona:

   ---
   agentMode: technical
   personaType: developer
   expertiseLevel: senior
   communicationStyle: technical
   ---
   
   # Persona: [User-Specified Role]
   
   (Generated from intake - requires completion)

   • Save to .github/ai-ley/personas/[category]/[name].md
   • Add to .ai-ley/shared/indexes/personas.md

   c. For Instructions:

   • Check if template exists at {{folders.templates.instructions}}/instruction-template.md
   • Create new instruction following template
   • Save to .github/ai-ley/instructions/[category]/[name].md
   • Add to .ai-ley/shared/indexes/instructions.md

   d. For Agents/Chatmodes:

   • Similar pattern using appropriate templates
   • Register in corresponding indexes

4. Record provenance:

   • Log creation in .project/GENERATED_CHANGES.md:

   ## Created During Intake ({{timestamp}})
   
   ### Tech Stack Entry
   
   - **Name**: Tailwind CSS
   - **Index**: `.ai-ley/shared/indexes/tech-stack.md`
   - **Reason**: User-specified during intake Q4

3.5: Append to INTAKE.md

After each question, append formatted response:

## Q4: Technology Stack

**Question**: What technologies will be used? (frontend, backend, database, infrastructure)

**Answer**:

- React (frontend framework)
- Node.js/Express (backend runtime/framework)
- PostgreSQL (database)
- Docker (containerization)
- Tailwind CSS (CSS framework) _[newly added to knowledge base]_

**Referenced Assets**:

- Tech stack entries: 5 items
- New entries created: 1 (Tailwind CSS)

**Answered**: 2025-11-09T12:45:23Z

---

Format Guidelines:

• Use consistent heading levels (## for questions)
• Include question text for context
• Format answers as lists, paragraphs, or structured data as appropriate
• Note any newly created assets with asterisk
• Include timestamp
• Add separator between questions

Step 4: Handle Uncertainty & Questions

If ambiguity or uncertainty occurs during intake:

1. Identify the uncertainty:

   • Unclear user response
   • Conflicting information
   • Missing critical context
   • Undefined capability referenced

2. Append to .ai-ley/shared/QUESTIONS.md:

## Q-INTAKE-001: Unclear Technology Stack Selection

**Context**: User mentioned "React Native" but also selected "Web (desktop)" as platform
**Question**: Is React Native intended for web or mobile? Or both?
**Impact**: Affects technology stack recommendations and project structure
**Source**: Intake session 2025-11-09T12:34:56Z
**Status**: Open
**Assigned To**: Project Lead
**Decision By**: 2025-11-16

**Options**:

1. React Native for mobile + React for web (separate codebases)
2. React Native for Web (unified codebase)
3. Clarify requirements with stakeholder

**Next Steps**:

- [ ] Clarify with user
- [ ] Update INTAKE.md with clarification
- [ ] Adjust technology stack accordingly

3. Continue with best guess or skip:
   • Document assumption in INTAKE.md
   • Flag for follow-up

Step 5: Completion & Requirements Generation

After the last question:

1. Mark intake as complete in .project/INTAKE.md:

---

## Intake Summary

**Status**: ✅ Complete
**Completed**: 2025-11-09T13:15:42Z
**Duration**: 41 minutes
**Questions Answered**: 10/10
**Assets Created**: 3 (1 tech stack, 1 persona, 1 instruction)
**Open Questions**: 1 (see QUESTIONS.md)

---

2. Prompt user for requirements generation:

✅ Intake complete!

Summary:
- 10 questions answered
- 3 new assets created
- 1 open question for follow-up

Would you like to generate or update `.project/REQUIREMENTS.md` using the `/requirements` prompt?

This will use your intake responses to create structured requirements.

Options:
1. Yes - Generate REQUIREMENTS.md from intake
2. No - I'll create requirements manually later
3. Preview - Show what requirements would be generated

Your choice (1/2/3):

3. If user chooses "Yes" or "Preview":

   a. Check if REQUIREMENTS.md exists:

   b. If REQUIREMENTS.md exists:

   ⚠️  .project/REQUIREMENTS.md already exists.
   
   Options:
   1. Update existing file (merge with intake data)
   2. Create REQUIREMENTS.generated.md (preserve existing)
   3. Show diff first
   
   Your choice (1/2/3):
   

   c. Trigger /requirements command:

   • Pass .project/INTAKE.md as input basis
   • Use mode=apply (or mode=preview if user selected preview)
   • Generate structured requirements from intake responses

   d. If creating REQUIREMENTS.generated.md:

   • Write to sibling file
   • Generate diff summary:

   # Requirements Diff Summary
   
   ## Additions (from intake)
   
   - Functional requirements based on project goals
   - Non-functional requirements inferred from constraints
   - Technology stack requirements from Q4
   
   ## Modifications
   
   ~ Updated team composition section
   ~ Added budget constraints
   
   ## Preserved (from existing)
   
   = Previous requirements R-001 through R-015

Step 6: Provenance & Documentation

Update .project/GENERATED_CHANGES.md:

### Files Created/Updated

1. **`.project/INTAKE.md`**

   - Session: {{type}} intake
   - Questions: 10/10 answered
   - Status: Complete

2. **`.ai-ley/shared/indexes/tech-stack.md`**

   - Added: Tailwind CSS
   - Reason: User-specified during intake

3. **`.github/ai-ley/personas/developer/tailwind-specialist.md`**

   - Created: New persona for Tailwind CSS expertise
   - Template: persona-template.md
   - Status: Generated (requires review)

4. **`.github/ai-ley/instructions/frontend/tailwind-best-practices.md`**
   - Created: Tailwind CSS best practices
   - Template: instruction-template.md
   - Status: Generated (requires completion)

### Assets Referenced

- Existing personas: 5
- Existing instructions: 8
- Existing tech stack items: 15

### Next Steps

- [ ] Review and complete generated persona: tailwind-specialist.md
- [ ] Complete instruction: tailwind-best-practices.md
- [ ] Resolve open question Q-INTAKE-001
- [ ] Generate REQUIREMENTS.md from intake (if selected)

Add metadata header to all generated files:

<!-- generated-by: ai-ley intake | timestamp: 2025-11-09T13:15:42Z | template:{{type}} -->

Step 7: Idempotency & Safety

Ensure idempotent behavior:

• Never delete user-provided content without explicit confirmation
• Append-only to INTAKE.md (never overwrite previous sessions)
• Use -generated suffix for updated files to preserve originals
• Track all changes in GENERATED_CHANGES.md
• Confirm before creating assets: Always ask before generating personas, instructions, etc.

Safety Checks:

• INTAKE.md preserves all previous content
• No hand-edited files overwritten
• All new assets use templates
• Indexes updated atomically
• Provenance documented
• Open questions captured in QUESTIONS.md

Quality Gates

Intake Session Quality:

• All required questions answered
• Responses formatted consistently
• Referenced assets identified and created/linked
• Provenance tracked for all generated assets
• Open questions documented
• Session marked as complete

Asset Creation Quality:

• Templates used for all new assets
• YAML frontmatter valid and complete
• File naming follows AI-LEY conventions
• Indexes updated with alphabetical ordering
• Generated file headers included

Requirements Generation Quality (if applicable):

• REQUIREMENTS.md generated from intake
• Existing requirements preserved (if any)
• Diff summary provided for review
• Structured requirements aligned with intake responses

Success Metrics

Intake Success:

• All questions answered or intentionally skipped
• User responses captured accurately
• Required assets created successfully
• Session completed without errors

Knowledge Base Enhancement:

• New assets properly registered in indexes
• Templates used consistently
• Provenance tracked for traceability
• Quality standards maintained

Requirements Alignment:

• Requirements generated align with intake
• No information loss from intake to requirements
• Structured and actionable requirements

Example 1: Basic Web App Intake

Command:

intake type=web-app

Expected Flow:

Loading template: .ai-ley/shared/templates/intake/web-app.md

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Question 1 of 10: Project Name
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

What is the name of your project?

Type: text
Required: Yes

> TaskMaster Pro

✅ Recorded

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Question 2 of 10: Project Description
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Provide a brief description of the project (1-2 paragraphs).

Type: textarea
Required: Yes

> A modern task management application for teams...
> [continues]

✅ Recorded

[... continues through all questions ...]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
✅ Intake Complete
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Summary:
- 10 questions answered
- 2 new assets created (1 persona, 1 instruction)
- 0 open questions

Would you like to generate .project/REQUIREMENTS.md? (y/n)

> y

Generating requirements from intake...
✅ Created .project/REQUIREMENTS.md

Next steps:
1. Review INTAKE.md for accuracy
2. Validate REQUIREMENTS.md completeness
3. Complete generated assets (see GENERATED_CHANGES.md)

Example 2: API Project Intake with Missing Template

Command:

intake type=graphql-api

Expected Flow:

⚠️  Template not found: .ai-ley/shared/templates/intake/graphql-api.md

Would you like to create a new intake template for "graphql-api"?

Options:
1. Create from base skeleton (recommended)
2. Use "api" template instead
3. Cancel and choose different template

Your choice (1/2/3): 1

Creating new template from skeleton...
✅ Created .ai-ley/shared/templates/intake/graphql-api.md
✅ Registered in .ai-ley/shared/indexes/templates.md

Loading template...

[... continues with questions ...]

Example 3: Intake with Custom Asset Creation

Command:

intake type=general

During Q6 (Team Composition):

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Question 6 of 10: Team Composition
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

What roles and expertise are available?

Suggested personas (from AI-Ley knowledge base):
  1. Senior Developer
  2. Backend Developer
  3. Frontend Developer
  4. DevOps Engineer
  5. QA Engineer

Enter numbers (comma-separated), or type custom roles:

> 1, 3, GraphQL API Specialist

⚠️  "GraphQL API Specialist" is not in the persona knowledge base.

Would you like to create it?
1. Yes - Create persona from template
2. No - Just record as freeform text

Your choice (1/2): 1

Creating persona: GraphQL API Specialist...
✅ Created .github/ai-ley/personas/developer/graphql-api-specialist.md
✅ Registered in personas index

Recorded:
- Senior Developer (existing)
- Frontend Developer (existing)
- GraphQL API Specialist (newly created)

Important Considerations

• Interactive Experience: Maintain conversational, helpful tone throughout
• Flexibility: Allow users to skip, go back, or modify responses
• Suggestions First: Always offer indexed options before accepting custom input
• Create Mindfully: Confirm before creating new assets; avoid bloat
• Preserve Context: Never lose user input; append-only approach
• Document Uncertainty: Capture ambiguities in QUESTIONS.md immediately

Best Practices

• Review Before Requirements: Always review INTAKE.md before generating REQUIREMENTS.md
• Complete Generated Assets: Review and complete stub personas/instructions promptly
• Update Indexes Regularly: Keep indexes current as knowledge base grows
• Use Appropriate Templates: Match intake template to project type for relevant questions
• Iterate: Allow users to run intake multiple times to refine responses

Common Pitfalls to Avoid

• Don't Overwhelm: Keep question sets focused (10-15 questions max)
• Don't Skip Validation: Validate required fields; guide users to complete answers
• Don't Create Duplicates: Check indexes before creating new assets
• Don't Lose Data: Always preserve previous intake sessions when appending
• Don't Auto-Generate Without Consent: Always ask before creating requirements

Extensibility

• Custom Templates: Users can create domain-specific intake templates
• Template Inheritance: Templates can reference base templates for common questions
• Multi-Stage Intake: Support progressive intake over multiple sessions
• Integration: Intake can feed other commands (plan, design, architecture)

Issue: Template not found

Solution: Offer to create from skeleton or use alternative template

Issue: Index file missing

Solution: Create index file from template; populate with intake-generated items

Issue: User provides ambiguous response

Solution: Ask clarifying question; log uncertainty in QUESTIONS.md

Issue: Requirements generation fails

Solution: Save intake; allow manual requirements creation; log error

Issue: Asset creation template missing

Solution: Generate minimal version; flag for review and completion

version: 1.0.0 updated: 2026-01-11 reviewed: 2026-01-11 score: 4.0

version: 1.0.0 updated: 2026-01-30 reviewed: 2026-01-30 score: 4