ldy26098
/
BMAD-METHOD
mirror of https://github.com/bmad-code-org/BMAD-METHOD.git
1
0
Fork 0
Code Packages Projects Releases Wiki Activity
BMAD-METHOD/expansion-packs/story-implementation/tasks/capture-learning-triage.md

5.9 KiB
Raw Blame History

Capture Learning Triage

Task Overview

Agent: architect
Action Type: learning-triage
Duration: 10-15 minutes
LLM-Optimized: Token-efficient structured capture

Purpose

Systematically capture and triage learnings from story implementation to drive continuous improvement and feed future epics.

Inputs

  • Story implementation file (docs/stories/epic{epic_number}.story{story_number}.story.md)
  • All review feedback from Round 1 reviews
  • Implementation fixes and changes
  • Quality gate results and metrics

Outputs

  • Learning items captured in story file under ## Learning Triage section
  • Categorized learning items with priorities and owners
  • Action items for immediate and future implementation

Learning Categories

ARCH_CHANGE (Architecture Changes Required)

  • Purpose: Technical debt or architecture improvements identified
  • Token Limit: 50 tokens per item
  • Format: ARCH: [Component] - [Issue] - [Impact] - [Owner: architect]
  • Priority: HIGH/MEDIUM/LOW
  • Timeline: Current epic / Next epic / Technical debt backlog

FUTURE_EPIC (Epic Candidate Features)

  • Purpose: Features or capabilities that emerged during implementation
  • Token Limit: 50 tokens per item
  • Format: EPIC: [Feature] - [Business Value] - [Complexity] - [Owner: po]
  • Priority: HIGH/MEDIUM/LOW
  • Timeline: Next sprint / Next quarter / Future roadmap

URGENT_FIX (Critical Issues Requiring Immediate Attention)

  • Purpose: Blockers or critical issues that need immediate resolution
  • Token Limit: 50 tokens per item
  • Format: URGENT: [Issue] - [Impact] - [Fix Required] - [Owner: dev/architect]
  • Priority: CRITICAL (resolve within current sprint)
  • Timeline: Immediate (within 1-2 days)

PROCESS_IMPROVEMENT (Development Process Enhancements)

  • Purpose: Workflow, tooling, or process improvements identified
  • Token Limit: 50 tokens per item
  • Format: PROCESS: [Area] - [Current State] - [Improvement] - [Owner: sm]
  • Priority: HIGH/MEDIUM/LOW
  • Timeline: Current sprint / Next sprint / Continuous improvement

TOOLING (Development Tooling and Infrastructure)

  • Purpose: Tools, automation, or infrastructure improvements needed
  • Token Limit: 50 tokens per item
  • Format: TOOLING: [Tool/System] - [Gap] - [Solution] - [Owner: infra-devops-platform]
  • Priority: HIGH/MEDIUM/LOW
  • Timeline: Current sprint / Next sprint / Infrastructure roadmap

KNOWLEDGE_GAP (Team Knowledge and Training Needs)

  • Purpose: Skills, knowledge, or training gaps identified during implementation
  • Token Limit: 50 tokens per item
  • Format: KNOWLEDGE: [Area] - [Gap] - [Training Need] - [Owner: sm/po]
  • Priority: HIGH/MEDIUM/LOW
  • Timeline: Current sprint / Next sprint / Long-term development

Execution Steps

Step 1: Review Implementation Context

CONTEXT_REVIEW:
- Story complexity: [SIMPLE/MODERATE/COMPLEX]
- Implementation time: [Actual vs Estimated]
- Quality gate failures: [Count and types]
- Review rounds required: [1/2/3+]
- Key technical challenges: [List top 3]

Step 2: Extract Learning Items

For each category, scan implementation evidence:

  • Review feedback patterns
  • Implementation fix patterns
  • Quality gate failure patterns
  • Time/effort variance patterns
  • Technical decision points

Step 3: Triage and Prioritize

TRIAGE_MATRIX:
High Priority: Blocks current/next sprint, affects team velocity
Medium Priority: Improves quality/efficiency, affects future work
Low Priority: Nice-to-have improvements, long-term optimization

Step 4: Assign Owners and Timelines

OWNERSHIP_ASSIGNMENT:
- architect: Architecture, technical debt, system design
- po: Business features, epic candidates, requirements
- dev: Implementation issues, code quality, technical fixes
- sm: Process improvements, team coordination, knowledge gaps  
- infra-devops-platform: Tooling, infrastructure, automation

Success Criteria

  • All learning categories reviewed and populated
  • Each item under 50 tokens with clear action owner
  • Priority and timeline assigned to each item
  • Immediate actions (URGENT_FIX) clearly identified
  • Future epic candidates captured with business value
  • Learning items added to story file under ## Learning Triage

Evidence Documentation

Update story file with:

## Learning Triage
**Architect:** [Name] | **Date:** [YYYY-MM-DD] | **Duration:** [X minutes]

### ARCH_CHANGE
- ARCH: [Component] - [Issue] - [Impact] - [Owner: architect] | Priority: [HIGH/MEDIUM/LOW] | Timeline: [Current/Next/Backlog]

### FUTURE_EPIC  
- EPIC: [Feature] - [Business Value] - [Complexity] - [Owner: po] | Priority: [HIGH/MEDIUM/LOW] | Timeline: [Next/Quarter/Future]

### URGENT_FIX
- URGENT: [Issue] - [Impact] - [Fix Required] - [Owner: dev/architect] | Priority: CRITICAL | Timeline: Immediate

### PROCESS_IMPROVEMENT
- PROCESS: [Area] - [Current State] - [Improvement] - [Owner: sm] | Priority: [HIGH/MEDIUM/LOW] | Timeline: [Current/Next/Continuous]

### TOOLING
- TOOLING: [Tool/System] - [Gap] - [Solution] - [Owner: infra-devops-platform] | Priority: [HIGH/MEDIUM/LOW] | Timeline: [Current/Next/Infrastructure]

### KNOWLEDGE_GAP
- KNOWLEDGE: [Area] - [Gap] - [Training Need] - [Owner: sm/po] | Priority: [HIGH/MEDIUM/LOW] | Timeline: [Current/Next/Long-term]

**Summary:** [X items captured] | [X urgent] | [X epic candidates] | [X process improvements]

Integration Points

  • Input from: validate_fixes (final architect review)
  • Output to: party-mode-learning-review (collaborative review)
  • Handoff: "Learning triage complete. Ready for collaborative review session."

LLM Optimization Notes

  • Token limits enforce brevity and focus
  • Structured format enables rapid scanning
  • Evidence-based categorization reduces subjective interpretation
  • Clear ownership prevents action item limbo
  • Timeline specificity enables proper backlog management