@cloudkinetix/bmad-enhanced/expansion-packs/ck-llm-agent-dev/workflows/prompt-optimization.yaml

Cloud-Kinetix enhanced fork of BMAD-METHOD - Breakthrough Method of Agile AI-driven Development with robust versioning and unified validation.

workflow:
  id: prompt-optimization
  name: Research-Driven Prompt Optimization
  description: Streamlined workflow for research-driven prompt optimization with adaptive planning and continuous improvement.
  type: optimization
  project_types:
    - prompt-improvement
    - token-optimization
    - quality-enhancement
    - safety-hardening
    - cost-reduction
    - performance-tuning
  approach: Research-driven optimization with adaptive phases based on current best practices and project-specific needs
  key_phases:
    research:
      description: Research current optimization techniques, tools, and methodologies
      agent: llm-engineer
      actions:
        - Research prompt optimization best practices and emerging techniques
        - Analyze current prompt performance and identify improvement opportunities
        - Investigate testing frameworks and evaluation methodologies
        - Study relevant case studies and optimization patterns
    optimize:
      description: Apply research findings to improve prompts systematically
      agent: llm-engineer
      actions:
        - Design improved prompts based on research findings
        - Implement optimization techniques appropriate for the use case
        - Apply current testing and validation methodologies
        - Document optimization rationale and approach
    validate:
      description: Test and validate optimization effectiveness
      agent: qa
      actions:
        - Execute comprehensive testing using research-backed methods
        - Validate improvements against established benchmarks
        - Verify safety and compliance requirements
        - Document results and lessons learned
    deploy:
      description: Deploy optimized prompts with monitoring
      agent: llm-engineer
      actions:
        - Research deployment best practices and implement gradual rollout
        - Set up monitoring and performance tracking
        - Validate production performance and user impact
        - Establish continuous improvement feedback loops
  success_criteria:
    - Research-informed optimization approach applied throughout
    - Measurable improvement in key performance metrics
    - Validated safety and quality standards maintained
    - Effective deployment with monitoring and feedback loops
    - Documented learnings for future optimization cycles
  notes: |
    This streamlined workflow emphasizes research-driven decision making over prescriptive steps.
    Teams should adapt phases and actions based on specific project needs and current best practices.