problem-solving

mway's avatarfrom mway

Structured problem-solving methodology. Activates for planning, design, debugging, or complex tasks. Use for systematic analysis and evidence-based reasoning.

⭐4stars🔀0forks📁View on GitHub🕐Updated Jan 8, 2026

When & Why to Use This Skill

This Claude skill implements a rigorous 5-phase problem-solving framework—Understand, Plan, Execute, Verify, and Reflect. Designed for complex, multi-step tasks, it ensures high-quality outcomes through systematic decomposition, evidence-based reasoning, and incremental validation. By following this structured methodology, the agent minimizes errors in planning and execution, making it an essential tool for technical design, debugging, and project management.

Use Cases

  • Complex Debugging: Systematically identifying root causes in large codebases by gathering evidence, isolating variables, and verifying fixes incrementally to prevent regressions.
  • Technical Architecture Design: Breaking down high-level system requirements into detailed technical specifications, identifying dependencies, and creating actionable implementation plans.
  • Multi-step Project Execution: Managing long-running technical migrations or feature developments with clear status tracking, parallel task identification, and rigorous sanity checks at every stage.
  • Evidence-Based Analysis: Conducting deep-dive investigations into system behavior or performance issues using a structured framework to ensure all constraints and unknowns are addressed.
nameproblem-solving
descriptionStructured problem-solving methodology. Activates for planning, design, debugging, or complex multi-step tasks. Use for systematic analysis and evidence-based reasoning.
allowed-tools["shell", "read_file", "apply_patch", "update_plan"]
short-descriptionSystematic problem-solving approach

Problem-Solving Methodology

Read these references:

  • references/methodology.md - 5-phase problem-solving framework (detailed)
  • references/behavior.md - Critical thinking and evidence-based reasoning
  • references/decomposition.md - Breaking problems into subproblems
  • references/parallelization.md - Identifying parallel execution opportunities

Instructions

Apply the 5-phase problem-solving framework from AGENT.md and references:

Phase 1: Understand (Decompose)

  • Gather all requirements and constraints
  • Identify unknowns and dependencies
  • Restate the problem to confirm understanding
  • Ask clarifying questions if needed
  • Read relevant code/documentation

Phase 2: Plan (Strategize)

  • Design approach from first principles
  • Make scope decisions (what's in/out)
  • Identify opportunities for parallelization
  • Break complex work into TODO items per task-management.md
  • Consider multiple solutions, pick best one

Phase 3: Execute (Implement)

  • Work through TODO list systematically
  • Update task status in real-time (exactly ONE in_progress)
  • Validate incrementally - don't save all verification for the end
  • Document decisions as you go
  • Apply critical thinking throughout

Phase 4: Verify (Validate)

  • Sanity check first: "Is this actually good/correct?"
  • Run mechanical verification (tests, lint, build)
  • Fix issues immediately, re-verify
  • Don't proceed until verification passes
  • Use evidence to confirm correctness

Phase 5: Reflect (Learn)

  • What worked well? What didn't?
  • Were estimates accurate?
  • Would a different approach have been better?
  • (This is internal reflection - share selectively)

All guidance from AGENT.md Task Tracking section also applies.