agile-project-mcp/docs/project-roles.md
2025-07-13 13:40:41 +01:00

5.1 KiB

Meet Your AI Development Team 👥

The Team Roster

Your AI development team consists of 8 specialised agents, each with their own personality and expertise:

👔 Product Owner (PO)

"I turn your ideas into actionable plans"

  • Takes your feature requests and breaks them down using MoSCoW prioritization (Must have, Should have, Could have, Won't have)
  • Manages the product backlog
  • Makes sure the team builds what you actually need

📊 Business Analyst (BA)

"I make requirements crystal clear"

  • Transforms high-level ideas into detailed user stories
  • Writes acceptance criteria that developers can actually implement
  • Ensures nothing gets lost in translation

📋 Project Manager (PM)

"I keep the trains running on time"

  • Organises work into manageable sprints
  • Assigns tasks to the right team members
  • Tracks progress and keeps you informed

🏃 Scrum Master (SM)

"I unblock the team and keep work flowing"

  • Monitors team progress every 3 minutes
  • Identifies and resolves blockers
  • Runs quick stand-ups to keep everyone aligned

💻 Developer

"I write the code that brings ideas to life"

  • Creates actual working code based on user stories
  • Produces clean, tested patches/PRs
  • Works on one story at a time for focused delivery

🧪 QA Engineer

"I make sure everything works as expected"

  • Tests features against acceptance criteria
  • Validates that code does what it should
  • Ensures quality before anything ships

🔍 Researcher

"I find answers when the team needs information"

  • Conducts quick research spikes when developers get stuck
  • Provides fact-based answers with sources
  • Time-boxed investigations to keep things moving

📝 Technical Writer

"I document everything clearly"

  • Creates user-facing documentation
  • Writes release notes
  • Keeps technical docs up to date

How They Work Together

A Day in the Life of Your AI Team

Let's say you request: "Build a user authentication system"

Hour 1: Planning

  1. 👔 PO receives your request and thinks: "This is a MUST have feature with several components"
  2. 📊 BA breaks it down:
    • "As a user, I want to sign up with email"
    • "As a user, I want to log in securely"
    • "As a user, I want to reset my password"

Hour 2: Sprint Kick-off 3. 📋 PM creates a sprint: "Let's tackle sign-up and log-in first" 4. 💻 Developers are assigned to different stories

Hour 3-4: Development 5. 💻 Dev #1 starts coding the sign-up flow 6. 🏃 SM checks in: "How's everyone doing?" 7. 💻 Dev #2 hits a snag: "What password hashing should we use?" 8. 🔍 Researcher jumps in: "Here's the current best practice..."

Hour 5: Testing 9. 🧪 QA tests the sign-up: "Works perfectly! Moving to 'Done'" 10. 📝 Tech Writer documents: "Here's how users sign up..."

Story Types They Handle

🚀 Feature Stories

New functionality for your project

  • Example: "Add shopping cart to e-commerce site"
  • Who's involved: BA → Dev → QA → Tech Writer

🐛 Bugfix Stories

Fixing issues in existing code

  • Example: "Log-in button doesn't work on mobile"
  • Who's involved: Dev → QA

🔬 Research Stories

Investigating unknowns

  • Example: "What's the best database for our needs?"
  • Who's involved: Researcher → Tech Writer

📚 Documentation Stories

Creating or updating docs

  • Example: "Write API documentation"
  • Who's involved: Tech Writer

The Workflow Magic

Sprint Rhythm

Every sprint follows this pattern:

  1. Planning (5-10 mins)

    • You approve what goes into the sprint
    • PM organizes the work
  2. Execution (30-40 mins)

    • Developers code
    • QA tests
    • SM keeps things moving
  3. Review (5 mins)

    • You see what was built
    • PO summarizes achievements

Automatic Check-ins

The Scrum Master runs "micro stand-ups" when:

  • 🕐 3 minutes pass without progress
  • 📊 5 tool calls happen without visible advancement
  • 🚧 Someone gets blocked

This keeps work flowing smoothly without your intervention!

Customising Your Team

Each agent can be customised via configuration:

developer:
  personality: "Senior developer with 10 years experience"
  model: "gpt-4"  # or any OpenAI-compatible model
  style: "Clean code advocate, loves TDD"

Want a more casual team? A more formal one? Different expertise levels? It's all configurable!

Best Practices

Getting Great Results

  1. Be specific - "Build a log-in page" is better than "add authentication"
  2. Trust the process - Let the team follow their workflow
  3. Review at checkpoints - Your input at sprint planning and review is valuable
  4. Let them iterate - The team will refine and improve with each sprint

When Things Get Stuck

  • The Scrum Master automatically detects and resolves most blocks
  • The Researcher provides answers when knowledge gaps appear
  • You can always intervene with new priorities or clarifications

Ready to Meet Your Team?

Start with a simple request and watch your AI team spring into action. They'll collaborate, solve problems, and deliver working software - just like a real development team, but available 24/7!