Skip to content

exercise-08-frd.md

Latest commit

 

History

History
60 lines (39 loc) · 2.49 KB

exercise-08-frd.md

File metadata and controls

60 lines (39 loc) · 2.49 KB

Exercise 08 — Generate the Functional Requirements Document

Duration: 4 minutes
Copilot Feature: FRD Custom Agent
Goal: Use the FRD agent to create testable user stories and use cases from the BRD and TSD.

Background

The FRD is the document development teams and QA engineers work from directly. It breaks down business requirements into use cases, user stories with Gherkin acceptance criteria, and a permissions matrix. The FRD Author agent synthesizes the BRD and TSD into this developer-ready document.

Step 1 — Switch to the FRD Agent

  1. In Copilot Chat, click the agent selector
  2. Select FRD Author

Step 2 — Send the FRD Generation Prompt

Copy and paste this prompt:

Read #file:requirement.md Create a comprehensive Functional Requirements Document saved as doc/frd.md.

Include:
- A User Roles & Permissions Matrix (Developer / Team Lead / Project Manager / QA Engineer vs all features)
- Detailed use cases for: Task Creation (UC-001), Task Assignment (UC-002), Task Dependency Management (UC-003), Task Status Tracking (UC-004), Task Listing and Filtering (UC-005), Project Progress Summary (UC-006)
- User stories for each use case with Given/When/Then acceptance criteria in Gherkin format
- A complete Functional Requirements Catalogue with FR-IDs linked to requirement sections
- Data validation rules for all input fields (Task ID, title, priority, status, dates)
- All notification triggers for task reassignment, dependency blocking, and status changes with recipient and content
- Error scenarios and user-facing error messages

Checkpoint — Docs Complete ✅

You should now have:

doc/
├── brd.md    ← Business requirements (from Exercise 06)
├── tsd.md    ← Technical architecture (from Exercise 07)
└── frd.md    ← Functional requirements + user stories (this exercise)

These three documents form the specification foundation for all remaining exercises. Every piece of code, database schema, test, and deployment script will trace back to one of these files.

Key Takeaway

Notice how each agent built on the previous output. BRD → TSD → FRD is the natural SDLC chain, and each agent was pre-wired to read the prior document. This is the power of document-driven agentic workflows — Copilot maintains context across the entire specification phase without you having to copy-paste content.

Next: Exercise 09 — Create Custom Instructions