Add OpenShift AI Virtual Team agents as source of truth

[jeremyeder]

Summary

This PR establishes the jeremyeder/vTeam repository as the source of truth for OpenShift AI Virtual Team agent definitions. It adds a comprehensive set of 17 specialized agents with rich personalities and domain expertise for realistic team simulation and collaboration.

New Directory Structure

Added:

• agents/ - Canonical location for vTeam agent definitions (17 agents)
• diagrams/ - Workflow diagrams and process documentation
• diagrams/ux-feature-workflow.md - UX feature development workflow diagram
• rhoai-ux-agents-vTeam.md - Comprehensive framework documentation

🔄 UX Feature Workflow Diagram

The diagrams/ux-feature-workflow.md contains a detailed mermaid flowchart showing how a UX feature flows through all 17 agents from ideation to sustaining engineering:

Workflow Phases:

1. 🚀 Ideation & Strategy - Parker, Aria, Ryan validate business opportunity
2. 🎨 Planning & Design - Uma, Felix, Steve, Casey create design strategy
3. 📋 Refinement & Breakdown - Derek, Olivia, Sam, Emma plan sprints
4. 🏗️ Architecture & Technical Planning - Archie, Stella, Lee, Phoenix design approach
5. ⚙️ Development & Implementation - Taylor, Tessa, Diego, Terry build and document
6. 🔄 Sustaining Engineering - Phoenix, Emma, Lee handle ongoing maintenance

Key Features:

• All 17 agents participate in realistic collaboration patterns
• 5 decision gates with proper validation and review processes
• Parallel work streams for development and documentation
• Agile ceremonies integrated throughout (facilitated by Sam)
• Feedback loops from production back to research and architecture
• Color-coded by role for visual clarity and organization

Agents Added (17 total)

Core Team Agents (3)

• agents/emma.md - Engineering Manager Agent (strategic, people-focused, protective of team wellbeing)
• agents/parker.md - Product Manager Agent (market-savvy, data-driven, business-focused)
• agents/taylor.md - Team Member Agent (pragmatic, detail-oriented, code quality champion)

Agile Role Agents (3)

• agents/sam.md - Scrum Master Agent (facilitator, process-oriented, diplomatically persistent)
• agents/olivia.md - Product Owner Agent (detail-focused, pragmatic negotiator, sprint guardian)
• agents/jack.md - Delivery Owner Agent (persistent tracker, cross-team networker, milestone-focused)

Engineering Role Agents (3)

• agents/archie.md - Architect Agent (visionary, systems thinker, long-term oriented)
• agents/stella.md - Staff Engineer Agent (technical authority, hands-on leader, mentoring focus)
• agents/lee.md - Team Lead Agent (technical coordinator, team advocate, execution-focused)

User Experience Agents (5)

• agents/aria.md - UX Architect Agent (holistic thinker, user advocate, ecosystem-aware)
• agents/uma.md - UX Team Lead Agent (design quality guardian, process driver, team coordinator)
• agents/felix.md - UX Feature Lead Agent (feature specialist, detail obsessed, accessibility champion)
• agents/steve.md - UX Designer Agent (creative problem solver, iteration enthusiast, user empathizer)
• agents/ryan.md - UX Researcher Agent (evidence seeker, insight translator, methodology expert)

Content Team Agents (4)

• agents/tessa.md - Technical Writing Manager Agent (quality-focused, deadline-aware, team coordinator)
• agents/diego.md - Documentation Program Manager Agent (timeline guardian, resource optimizer, dependency tracker)
• agents/casey.md - Content Strategist Agent (big picture thinker, standard setter, cross-functional bridge)
• agents/terry.md - Technical Writer Agent (user advocate, technical translator, accuracy obsessed)

Special Team Agent (1)

• agents/phoenix.md - PXE (Product Experience Engineering) Agent (customer impact predictor, risk assessor, lifecycle thinker)

Agent Features

Each agent includes:

• Rich Personalities: Distinct communication styles and behavioral patterns
• Role-Specific Competencies: Technical skills and domain expertise appropriate to their role
• OpenShift AI Platform Knowledge: Deep understanding of ML platform components, workflows, and user needs
• Signature Phrases: Authentic dialogue patterns for realistic interactions
• Appropriate Tool Access: Tool permissions aligned with role responsibilities

Design Philosophy

This agent system models realistic team dynamics by including:

1. Different Seniority Levels - Creates authentic interaction patterns and knowledge gradients
2. Cognitive Diversity - Different experience levels approach problems differently
3. Productive Tensions - Natural conflicts between roles surface important trade-offs
4. Role-Appropriate Communication - Different levels explain concepts with appropriate depth

Technical Stack Coverage

Agents have deep knowledge of the complete OpenDataHub ecosystem:

• Languages: Python, Go, JavaScript/TypeScript, Java, Shell/Bash
• ML/AI Frameworks: PyTorch, TensorFlow, XGBoost, Scikit-learn, HuggingFace, vLLM
• Container & Orchestration: Kubernetes, OpenShift, Docker, Podman, CRI-O
• ML Operations: KServe, Kubeflow, ModelMesh, MLflow, Ray, Feast
• Data Processing: Apache Spark, Argo Workflows, Tekton
• Monitoring: Prometheus, Grafana, OpenTelemetry

Source of Truth Establishment

This PR establishes jeremyeder/vTeam/agents/ as the canonical location for agent definitions, enabling:

• Reusability - Agents can be referenced by multiple vTeam projects
• Centralized Management - Single location for agent updates and maintenance
• Consistency - Shared agent framework across all vTeam implementations
• Version Control - Proper tracking of agent evolution and changes

Directory Organization

• agents/ - Focus on agent definitions and personalities
• diagrams/ - Process workflows and collaboration patterns
• Root level - Framework documentation and configuration

Integration Pattern

Other vTeam projects (like vTeam-console) should reference these agents:

# Example integration
agents_source: "jeremyeder/vTeam/agents/"  
diagrams_source: "jeremyeder/vTeam/diagrams/"
agents_format: "claude_code_agent_markdown"

Use Cases for Workflow Diagram

The diagrams/ux-feature-workflow.md can be used for:

• 🎓 Team Training - Understanding role relationships and handoff patterns
• 📋 Process Documentation - Standardizing feature development approaches
• 🤝 Stakeholder Communication - Explaining how features flow through the team
• 🔧 Tool Configuration - Setting up automation based on collaboration patterns
• 📊 Process Optimization - Identifying bottlenecks and improvement opportunities

Test Plan

• Verify all 17 agent files are properly formatted with YAML frontmatter
• Confirm agent personalities and competencies align with their roles
• Test that signature phrases and communication styles are distinct
• Validate OpenShift AI domain knowledge is accurate and comprehensive
• Ensure tool permissions are appropriate for each role
• Verify source documentation completeness
• Validate workflow diagram accuracy against agent interaction patterns
• Confirm all 17 agents are represented in the workflow
• NEW: Verify proper directory organization (agents/, diagrams/)

Benefits

1. Source of Truth - Single canonical location for agent definitions
2. Realistic Team Dynamics - Natural conflicts and alliances between roles
3. Deep Domain Expertise - OpenShift AI platform knowledge for relevant discussions
4. Rich Personalities - Distinct communication patterns and behavioral traits
5. Scalable Framework - Foundation for additional roles and specializations
6. Process Visualization - Clear understanding of how teams actually collaborate
7. Training Resources - Comprehensive workflow for onboarding and process improvement
8. 🆕 Organized Structure - Clean separation between agent definitions and process diagrams

🤖 Generated with Claude Code

[jeremyeder]

1. I committed the source for the agents as rhoai-ux-agents-vTeam.md. That file was derived from @dgutride google document.
2. I created mermaid diagrams that try to document training, process changes, stakeholder comms, and identifying bottlenecks.