@cloudkinetix/bmad-enhanced
Version:
Cloud-Kinetix enhanced fork of BMAD-METHOD - Breakthrough Method of Agile AI-driven Development with robust versioning and unified validation.
312 lines (268 loc) • 11.7 kB
YAML
name: Enhanced LLM-Native JIRA Workflows
description: Example workflows demonstrating advanced context management, prompt optimization, and multi-turn reasoning
version: 1.0.0
workflows:
- id: epic-breakdown-with-reasoning
name: Intelligent Epic Breakdown
description: Break down complex epics using multi-turn reasoning and context awareness
tags: [epic, planning, multi-turn]
context_template: feature_epic_breakdown
phases:
- name: Analysis
description: Analyze epic complexity with context awareness
steps:
- id: load-context
task: Execute context-manager
action: |
Load or create epic breakdown context
Check for previous breakdown attempts
Apply team's preferred patterns
- id: analyze-epic
task: Execute prompt-optimizer
action: |
Select optimal epic analysis prompt based on:
- Epic size and complexity
- Team's historical patterns
- Available context richness
Execute analysis with enhanced prompt
- id: present-analysis
task: Execute reasoning-engine
action: |
Present epic analysis with progressive disclosure:
- Summary for quick review
- Detailed breakdown on request
- Risk factors highlighted
- Suggest breakdown approaches
- name: Interactive Planning
description: Guide user through story generation with checkpoints
steps:
- id: breakdown-conversation
task: Execute interaction-flows
action: |
Start epic_breakdown_flow:
Turn 1: "I've analyzed epic {{epic_key}}. It's {{complexity}} with {{estimate}} potential stories."
Turn 2: Present breakdown options (vertical, horizontal, risk-based)
Turn 3: Generate stories based on chosen approach
Turn 4: Refine with user input
- id: checkpoint-progress
task: Execute checkpoint-manager
action: |
Create checkpoint after each turn
Allow resume if interrupted
Track decisions made
- name: Story Creation
description: Create stories with optimized prompts
steps:
- id: optimize-creation
task: Execute prompt-chains
action: |
Chain story creation prompts:
1. Validate story completeness
2. Check dependencies
3. Create in JIRA with proper links
4. Update BMAD documentation
- id: verify-sync
task: Execute universal-sync-analysis
action: |
Verify three-way sync health
Update context with created stories
Generate sync report
- id: sprint-planning-adaptive
name: Adaptive Sprint Planning
description: Plan sprints with intelligent capacity analysis and story selection
tags: [sprint, planning, adaptive]
context_template: sprint_planning_standard
phases:
- name: Capacity Intelligence
description: Smart capacity calculation with pattern learning
steps:
- id: analyze-patterns
task: Execute context-analyzer
action: |
Analyze team's sprint patterns:
- Historical velocity trends
- Completion rates by story type
- Common planning mistakes
Generate insights for current sprint
- id: calculate-capacity
task: Execute prompt-library
action: |
Use optimized capacity calculation prompt
Factor in:
- Team availability
- Historical accuracy
- Current context (holidays, releases)
- name: Story Selection
description: Intelligent story selection with multi-criteria optimization
steps:
- id: selection-conversation
task: Execute decision-trees
action: |
Navigate sprint_planning_decision_tree:
- Check capacity vs velocity
- Evaluate story readiness
- Consider dependencies
- Balance priorities
- id: optimize-selection
task: Execute prompt-optimizer
action: |
Use context-aware story selection:
- Team's skill distribution
- Epic priorities
- Technical dependencies
- Risk balance
- id: intelligent-sync-recovery
name: Sync Operation with Recovery
description: Demonstrate checkpoint recovery for interrupted sync operations
tags: [sync, recovery, resilience]
phases:
- name: Sync Preparation
description: Prepare for sync with checkpoint creation
steps:
- id: create-sync-checkpoint
task: Execute checkpoint-manager
action: |
Create checkpoint before sync:
- Current sync scope
- Entity states
- User preferences
Mark as resumable
- name: Simulated Interruption
description: Handle sync interruption gracefully
steps:
- id: detect-incomplete
task: Execute reasoning-engine
action: |
On next session:
"I see you have an incomplete sync operation from {{time_ago}}.
You were syncing {{entity_count}} stories and completed {{completed_count}}.
Would you like to continue where you left off?"
- id: resume-sync
task: Execute checkpoint-manager
action: |
If user confirms:
- Restore full context
- Continue from last successful operation
- Show what was completed and what remains
- id: personalized-daily-standup
name: Context-Aware Daily Standup
description: Generate standups that learn from team patterns
tags: [standup, daily, personalized]
phases:
- name: Pattern Learning
description: Learn team's standup preferences
steps:
- id: analyze-preferences
task: Execute context-analyzer
action: |
Analyze past standup patterns:
- Preferred format and detail level
- Common discussion topics
- Time constraints
- Team distribution
- name: Intelligent Generation
description: Create personalized standup agenda
steps:
- id: optimize-standup-prompt
task: Execute prompt-optimizer
action: |
Select standup template based on:
- Day of week (Monday vs Friday)
- Sprint phase (early, mid, late)
- Recent team events
- Detected urgency
- id: progressive-standup
task: Execute interaction-flows
action: |
Use progressive disclosure:
- Quick summary first
- Detailed sections on request
- Focus on anomalies and risks
- Suggest discussion topics
- id: bug-investigation-assistant
name: Intelligent Bug Investigation
description: Multi-turn bug analysis with context-aware insights
tags: [bug, investigation, multi-turn]
phases:
- name: Bug Discovery
description: Smart bug identification and context gathering
steps:
- id: understand-issue
task: Execute reasoning-engine
action: |
Natural conversation:
User: "The login is broken"
Assistant: "I'll help investigate the login issue. Let me check a few things..."
Use context to understand:
- Component affected
- Recent changes
- Similar past issues
- name: Guided Investigation
description: Lead user through systematic investigation
steps:
- id: investigation-flow
task: Execute decision-trees
action: |
Use bug_investigation_tree:
- Check error patterns
- Review recent deployments
- Analyze related tickets
- Generate hypotheses
- id: adaptive-depth
task: Execute prompt-optimizer
action: |
Adjust investigation depth based on:
- User expertise level
- Issue complexity
- Available time
- Prior investigation success
example_usage:
epic_breakdown_scenario: |
User: "Break down the payment processing epic"
Assistant: [Initializes context, loads patterns]
"I'll help you break down the payment processing epic. Let me analyze it first..."
[Uses optimized prompt for epic analysis]
"I've analyzed PAYMENT-100: 'Payment Processing System'. Here's what I found:
📊 **Epic Analysis**
- Estimated size: Large (8-12 stories)
- Technical complexity: High (external integrations)
- Risk factors: PCI compliance, third-party dependencies
- Team experience: Medium (new payment provider)
How would you like to break this down?
1. **By payment flow** (vertical slices) - Recommended
2. **By technical layer** (frontend/backend/integration)
3. **By risk** (tackle unknowns first)
4. **Custom approach**"
User: "Let's go with payment flow"
Assistant: [Creates checkpoint, continues conversation]
"Good choice! Breaking down by payment flow will give us complete user journeys.
Based on the requirements, I suggest these payment flows:
1. **Credit Card Processing** (3 stories)
2. **PayPal Integration** (2 stories)
3. **Refund Management** (2 stories)
4. **Payment History** (1 story)
5. **Error Handling & Recovery** (2 stories)
Would you like me to generate the detailed stories for these flows?"
[Continues with story generation, creating checkpoints at each step]
sync_recovery_scenario: |
[Session 1]
User: "Sync all stories in the current sprint"
Assistant: "I'll sync all stories in Sprint 42. Found 15 stories to sync..."
[Creates checkpoint, begins sync]
[Network interruption after 8 stories]
[Session 2 - Next day]
Assistant: "Welcome back! I see you have an incomplete sync operation from yesterday.
You were syncing 15 stories in Sprint 42 and completed 8 successfully.
✅ Completed: AUTH-101, AUTH-102, PAY-201...
⏳ Remaining: PAY-203, UI-301, UI-302...
Would you like to continue syncing the remaining 7 stories?"
User: "Yes, continue"
Assistant: [Restores context, resumes operation]
"Resuming sync from PAY-203..."
key_benefits:
- context_awareness: Operations remember previous interactions and learn patterns
- intelligent_prompting: Prompts are optimized based on context and performance
- natural_conversations: Complex operations feel like helpful discussions
- resilient_operations: Interruptions are handled gracefully with checkpoints
- continuous_improvement: System learns and adapts while preserving privacy