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/.github/ISSUE_TEMPLATE/agent_improvement.md

6.1 KiB
Raw Blame History

name about title labels assignees
🎭 Agent Improvement Suggest improvements to specific BMAD Claude agents [AGENT] [Agent Name] -
agent-improvement
enhancement

🎭 Agent Improvement

📊 Target Agent

  • 📊 Analyst (Mary) - Market research, competitive analysis, project briefing
  • 🏗️ Architect (Winston) - System design, technical architecture
  • 👨‍💻 Dev (James) - Implementation, coding, technical development
  • 📋 PM (John) - Project management, planning, coordination
  • 🔍 QA (Quinn) - Quality assurance, testing, validation
  • 🎯 Scrum Master (Bob) - Agile process management, team facilitation
  • 🔄 Cross-Agent - Multi-agent workflows and collaboration

🎯 Improvement Type

  • 🧠 Persona Enhancement - Improve agent character/personality
  • 📝 Prompt Optimization - Better system prompts and instructions
  • 🛠️ New Capabilities - Add new commands or skills
  • 📚 BMAD Integration - Better methodology adherence
  • 🤝 Collaboration - Improve cross-agent workflows
  • 💾 Memory/Context - Better context management
  • 📊 Output Quality - Improve response quality
  • 🎨 User Experience - Better interaction patterns

📋 Current Behavior

What does the agent do now?

Current Strengths

Current Limitations

Example Current Response

<!-- Paste an example of current agent behavior -->

✨ Proposed Improvement

What should the agent do differently?

Detailed Description

Expected Benefits

Example Improved Response

<!-- Show what the improved response would look like -->

🧪 Testing Scenario

Test Prompt

<!-- Provide a specific prompt to test the improvement -->

Success Criteria

  • Criterion 1
  • Criterion 2
  • Criterion 3

Evaluation Method

  • Manual testing in Claude Code
  • Comparison with previous responses
  • User feedback collection
  • Automated scoring (if applicable)

🎭 BMAD Methodology Alignment

BMAD Integration

  • Enhances existing BMAD workflow adherence
  • Adds new BMAD artifact usage
  • Improves BMAD template integration
  • Better cross-agent BMAD coordination
  • Maintains BMAD methodology integrity

Agent Role Consistency

  • Stays true to agent's primary BMAD function
  • Enhances domain expertise appropriately
  • Maintains professional persona
  • Supports BMAD workflow patterns

📈 Impact Assessment

User Experience Impact

  • 🟢 Significantly improves user experience
  • 🟡 Moderately improves user experience
  • 🟠 Minor improvement
  • 🔴 Minimal impact

Agent Performance Impact

  • 🟢 Major improvement in agent capability
  • 🟡 Moderate improvement
  • 🟠 Small improvement
  • 🔴 Minimal improvement

Implementation Complexity

  • 🟢 Simple - Prompt/template changes only
  • 🟡 Medium - Some code changes required
  • 🟠 Complex - Significant changes needed
  • 🔴 Major - Architectural changes required

🔧 Implementation Details

Prompt Changes Required

New Dependencies

  • New task files in bmad-core/tasks/
  • New templates in bmad-core/templates/
  • New data files in bmad-core/data/
  • Agent template modifications
  • Build system changes

Tool/Command Changes

  • New BMAD commands needed
  • Existing command modifications
  • New tool permissions required
  • Memory/persistence changes

📊 Comparison Analysis

Similar Functionality

Best Practices

Differentiation

🔗 Related Work

  • Related to #(issue number)
  • Inspired by #(issue number)
  • Builds on #(issue number)

External References

  • BMAD methodology documentation
  • Claude Code best practices
  • Agent prompt engineering guides

📋 Additional Context

User Feedback

Research/Evidence

Examples from Other Domains

🤝 Implementation Plan

Phase 1: Planning

  • Finalize improvement specification
  • Design prompt modifications
  • Plan testing approach

Phase 2: Implementation

  • Modify agent templates/prompts
  • Add any new dependencies
  • Update build system if needed

Phase 3: Testing

  • Test with sample scenarios
  • Validate BMAD integration
  • Collect feedback

Phase 4: Refinement

  • Iterate based on testing
  • Document changes
  • Update examples

🤝 Contribution

  • I can implement the prompt changes
  • I can help with testing and validation
  • I can provide additional use cases
  • I can help with documentation
  • I need guidance but want to contribute

Please ensure you've:

  • Clearly identified the target agent(s)
  • Provided specific examples of current vs. improved behavior
  • Considered BMAD methodology alignment
  • Thought through implementation complexity
  • Included testing criteria