/tech-req

name: tech-req
description: Technical requirements gathering - Evaluate technologies and architecture through focused questions and guided questionnaire
version: 1.0.0
argument-hint: '[existing-prd-file-or-project-name] [--process TECH_REQ_QUESTIONS.md]'

Tech-Req Command - Technical Requirements Discovery

You are facilitating **TECHNICAL REQUIREMENTS DISCOVERY** using a hybrid approach: **Interactive critical questions** followed by a **self-paced questionnaire**.

⚠️ **IMPORTANT WORKFLOW**:

1. **Read PRD context** (if exists) to extract constraints 2. **Ask 2-3 critical questions** interactively (architecture, cloud, framework category) 3. **Generate questionnaire file** (TECH_REQ_QUESTIONS.md) for remaining decisions 4. **User fills questionnaire** at their own pace 5. **Process questionnaire** (`/tech-req --process TECH_REQ_QUESTIONS.md`) to generate TECH_REQ.md

⚠️ **KEY PRINCIPLES**:

• **Smart Filtering**: Show only 2-3 most relevant options per decision (filtered by PRD constraints)
• **Sequential Decisions**: Later questions depend on earlier answers
• **No Information Overload**: Never present 4+ options simultaneously
• **Strict Separation**: PRD = what/why/who, Tech-Req = how/with what

Initial Input

$ARGUMENTS

If `--process TECH_REQ_QUESTIONS.md` flag detected:

• Skip to [Questionnaire Processing](#questionnaire-processing) section
• Read answers from file and generate TECH_REQ.md

Otherwise, proceed with interactive discovery.

---

Step 1: Extract PRD Context

If PRD.md exists, read it and extract:

• **Deployment Preference**: Cloud (AWS/Azure/GCP) / Local / Hybrid
• **Scale**: Expected users/traffic
• **Budget**: Infrastructure cost constraints
• **Team Technical Comfort**: Beginner / Intermediate / Advanced
• **Must Use/Avoid**: Required or forbidden technologies
• **Performance Requirements**: Response time targets
• **Data Needs**: Storage type and volume

**Document Extracted Constraints**:

## PRD Constraints Extracted

- **Deployment**: [Cloud provider or preference]
- **Scale**: [User count, traffic estimates]
- **Budget**: [Monthly infrastructure budget]
- **Team Comfort**: [Technical skill level]
- **Must Use**: [Required technologies]
- **Must Avoid**: [Technologies to avoid]
- **Performance**: [Targets if specified]
- **Data**: [Storage needs]

If no PRD exists, ask user: "I don't see a PRD.md file. Would you like me to:

1. Ask a few constraint questions now 2. Generate a PRD first using `/prd` 3. Proceed with general recommendations (no filtering)"

---

Step 2: Interactive Critical Questions (2-3 questions only)

Ask **ONE question at a time**, wait for answer, adapt next question.

Question 1: Architecture Pattern (ALWAYS ASK)

**Apply Smart Filtering** based on PRD constraints:

**Filtering Logic**:

• If "single author blog" or "documentation site" → Show only SSG + JAMstack (skip SSR, microservices)
• If "< 10 users" or "internal tool" → Show only Monolithic + Serverless (skip microservices)
• If "need real-time features" → Show SSR + Serverless (skip pure SSG)
• If "> 100 users" and "multiple teams" → Show Microservices + Hybrid
• **Default** (no constraints): Show Monolithic + Serverless + JAMstack (skip microservices unless scale justifies)

**Present Only 2-3 Filtered Options**:

"Based on your requirements [cite PRD constraints], here are the most suitable architecture patterns:

---

Option 1: [Most Suitable Architecture for Constraints]

**What it is**: [One sentence description]

**Pros**:

• ✅ [Key advantage 1]
• ✅ [Key advantage 2]
• ✅ [Key advantage 3]

**Cons**:

• ❌ [Key limitation 1]
• ❌ [Key limitation 2]

**Best for**: [Specific use case] **Matches your**: [How it aligns with PRD constraints] **Cost**: [Monthly estimate]

---

Option 2: [Second Most Suitable Architecture]

[Same structure as Option 1...]

---

[Optional Option 3 only if legitimately applicable]

---

**My Recommendation**: [Architecture] because [2-3 specific reasons tied to PRD constraints].

**Question**: Which architecture pattern aligns with your vision?"

**Wait for user answer before proceeding.**

---

Question 2: Cloud Provider (CONDITIONAL)

**Only ask if**:

• PRD deployment preference is "Cloud"
• PRD didn't specify cloud provider preference

**Skip if**:

• PRD says "AWS only" → Auto-select AWS, document decision
• Deployment is Local/Hybrid → Skip to framework question

**Apply Smart Filtering**:

• If Budget < $10/month → Skip Azure/GCP, show AWS (best free tier)
• If "Must Use: Microsoft stack" → Show only Azure
• If no constraints → Show AWS + 1 alternative (GCP if focus is data/ML, Vercel if SSG architecture chosen)

**Present Only 2 Filtered Options**:

"For cloud hosting, here are your best options based on [architecture choice] and [budget/constraints]:

---

Option 1: AWS

**Why for you**: [Specific reason based on architecture + constraints] **Monthly Cost**: ~$[X] (for your scale) **Key Services**: [3-4 services you'd use]

---

Option 2: [GCP / Azure / PaaS]

**Why for you**: [Specific reason] **Monthly Cost**: ~$[X] **Key Services**: [3-4 services you'd use]

---

**My Recommendation**: [Provider] because [reason tied to your constraints].

**Question**: Which cloud provider do you prefer?"

**Wait for user answer before proceeding.**

---

Question 3: Framework Category (CONDITIONAL)

**Question depends on architecture answer**:

**If SSG architecture chosen** → Ask about static site generators **If SSR/Monolithic chosen** → Ask about full-stack frameworks **If Serverless chosen** → Ask about serverless framework preferences

**Apply Smart Filtering** (example for SSG):

• If "performance critical" → Show Astro + Hugo (skip heavier options)
• If "familiar with React" → Show Next.js + Astro
• If "beginner" → Show Hugo + 11ty (skip React-based)
• **Always limit to 2-3 options**

**Example for SSG Architecture**:

"For static site generation with [your requirements], here are the b