Skip to content

v1.1 - Discovery-First Development

Released: November 2025

This release transforms DevEnv into a discovery-first development system where requirements extraction precedes all implementation.


Highlights

  • New Skill: discovery-interviewer - 90-minute structured requirements interviews
  • New Skill: test-generator - BDD/Gherkin test scenario generation
  • Updated: saas-project-orchestrator now begins with discovery as Phase 0
  • New Methodology: Scenarios serve as requirements, tests, and documentation

New: Discovery Interviewer Skill

The discovery-interviewer skill conducts comprehensive 90-minute structured interviews that extract complete requirements before any code is written.

What It Does

Guides you through 8 sections in 90 minutes:

Section Duration Focus
Product Vision 10 min Elevator pitch, value proposition, business model
User Personas 10 min User types, goals, pain points, journeys
Core Workflows 20 min Step-by-step actions, data, decisions, errors
Data & Entities 15 min Things the system manages, relationships
Edge Cases 10 min Boundaries, concurrent operations, deletions
Success Criteria 10 min Acceptance criteria, performance targets
Constraints 10 min Tech requirements, compliance, integrations
Review 5 min Completeness check, prioritization

Dual Persona Mode

The skill operates in two modes:

  1. Interviewer Mode: When you know the answer, it extracts and documents
  2. Product Expert Mode: When you don't know, it provides industry best practices and guides you to decisions

Output

A 15-25 page discovery document containing:

  • Complete workflows with every step documented
  • Full data model with entity definitions
  • Edge cases and boundary conditions
  • Success criteria and acceptance definitions
  • Technical constraints and integration requirements

New: Test Generator Skill

The test-generator skill creates comprehensive BDD/Gherkin test suites from discovery output.

Three-Layer Testing

Layer Coverage Focus
API Tests 40% Backend endpoints, business logic
UI Tests 45% Components, forms, interactions
E2E Tests 15% Complete user workflows

What It Creates

  • Gherkin .feature files organized by layer
  • TypeScript step definitions
  • Cucumber configuration
  • Traceability matrix (task to test mapping)

Tag System

@api @crud @smoke
Scenario: Create project with valid data
  Given I have a valid project payload
  When I send a POST request to "/api/projects"
  Then the response status should be 201

Tags organize tests by:

  • Layer: @api, @ui, @e2e
  • Priority: @smoke, @critical, @regression
  • Type: @positive, @negative, @edge-case, @security

Updated Workflow

Before v1.1

Orchestrator → Brief → Schema → Backend → Frontend → Features

After v1.1

Phase 0: discovery-interviewer → Complete discovery doc
   |
Phase 1: test-generator → Gherkin scenarios from workflows
   |
Phase 2: schema-architect → Extract from discovery + scenarios
   |
Phase 3: product-brief-writer → Synthesize all artifacts
   |
Phase 4: roadmap-planner → Phase scenarios into delivery waves
   |
Phase 5+: Implementation (backend, frontend, features)

Why This Order?

  1. Discovery First: Quality here determines quality everywhere
  2. Scenarios Second: Workflows become both acceptance criteria AND automated tests
  3. Schema Third: Extracted from discovery entities + scenario data needs
  4. Brief Fourth: Synthesizes and references other artifacts
  5. Roadmap Fifth: Groups scenarios by priority into delivery phases

Triple-Duty Gherkin

Gherkin scenarios now serve three purposes simultaneously:

1. Requirements Specification

Feature: Team Invitations

  Scenario: Admin invites new team member
    Given I am an admin of organization "Acme Corp"
    When I invite "newuser@example.com" to the team
    Then an invitation email should be sent
    And the invitation should expire in 7 days

This documents: who can invite (admins), what happens (email sent), business rules (7-day expiration).

2. Acceptance Tests

The same scenario becomes an automated test that validates implementation.

3. Living Documentation

Scenarios serve as always-current documentation because they're verified by tests.


Coverage Standards

Feature Type Minimum Tests API UI E2E
CRUD Entity 15-20 8-10 5-7 2-3
Authentication 17-22 6-8 8-10 3-4
Payment Flow 15-19 5-6 6-8 4-5

Getting Started

Start Discovery

"I want to build a SaaS for [your idea]"

The orchestrator automatically calls discovery-interviewer as Phase 0.

Generate Tests

"Generate tests for [feature name]"

The test-generator creates Gherkin scenarios with full coverage.


Benefits

For Builders

  • Clear target with no ambiguity
  • Test-driven with automated validation
  • Less rework from correct requirements upfront

For Quality

  • Complete requirements lead to complete implementation
  • Edge cases handled from the start
  • 70-80% test coverage standard

For Business

  • Build it right once instead of iterating forever
  • Clear scope prevents endless scope creep
  • Validated assumptions early in the process