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/bmad-agent/tasks/udtm_task.md

2.9 KiB
Raw Blame History

Ultra-Deep Thinking Mode (UDTM) Task

Purpose

Execute rigorous analysis and verification protocol to ensure highest quality decision-making and implementation. Store UDTM analysis results at .ai/quality/validations/udtm-analysis-{task}-{date}.md.

Protocol

Phase 1: Multi-Angle Analysis (30 minutes minimum)

  • Technical Perspective: Correctness, performance, maintainability
  • Business Logic Perspective: Alignment with requirements
  • Integration Perspective: Compatibility with existing systems
  • Edge Case Perspective: Boundary conditions and failure modes
  • Security Perspective: Vulnerabilities and attack vectors
  • Performance Perspective: Resource usage and scalability

Phase 2: Assumption Challenge (15 minutes)

  1. List all assumptions made during analysis
  2. Challenge each assumption - attempt to disprove
  3. Document evidence for/against each assumption
  4. Identify critical dependencies on assumptions

Phase 3: Triple Verification (20 minutes)

  • Source 1: Official documentation/specifications
  • Source 2: Existing codebase patterns and standards
  • Source 3: External validation (tests, tools, references)
  • Cross-reference: Ensure all three sources align

Phase 4: Weakness Hunting (15 minutes)

  • What could break this solution?
  • What edge cases might we have missed?
  • What are the failure modes?
  • What assumptions are we making that could be wrong?
  • What integration points could fail?

Phase 5: Final Reflection (10 minutes)

  • Re-examine entire reasoning chain from scratch
  • Document confidence level (must be >95% to proceed)
  • Identify any remaining uncertainties
  • Confirm all quality gates can be met

Output Requirements

Document all phases with specific findings, evidence, and confidence assessments. Save the complete UDTM analysis at .ai/quality/validations/udtm-analysis-{task}-{date}.md.

Success Criteria

  • All phases completed with documented evidence
  • Confidence level >95%
  • All assumptions validated or flagged as risks
  • Quality gates confirmed achievable

Usage Instructions

  1. Execute this task before any major implementation or decision
  2. Document all findings in the UDTM Analysis Template at .ai/quality/validations/udtm-analysis-{task}-{date}.md
  3. Do not proceed without achieving >95% confidence
  4. Share analysis with team for brotherhood review
  5. Store completed analysis for future reference and learning

Integration with BMAD Workflow

  • BREAK Phase: Use UDTM for problem decomposition
  • MAKE Phase: Apply before each implementation sprint
  • ANALYZE Phase: Execute for issue investigation
  • DELIVER Phase: Final validation before deployment
  • Documentation: All UDTM analyses stored at .ai/quality/validations/ for tracking and learning
  • Quality Metrics: UDTM completion tracked in project quality metrics