---

name: solving-problems description: AI agent applies a 5-phase systematic framework for tackling complex problems when conventional approaches fail. Use when stuck, blocked, or troubleshooting issues. category: methodology triggers:

• stuck
• blocked
• problem solving
• troubleshooting
• can't figure out

---

Solving Problems

Purpose

Apply a systematic 5-phase framework for complex problems:

• Define problem boundaries clearly before investigating
• Generate multiple hypotheses before testing any
• Test hypotheses efficiently with time-boxing
• Fix root causes, not just symptoms
• Prevent recurrence with process improvements

Quick Start

1. Define (5-10 min) - Clarify what IS and IS NOT the problem, set success criteria
2. Hypothesize (10-15 min) - Generate 3+ hypotheses BEFORE testing any
3. Test (time-boxed) - Test cheapest hypothesis first, mark confirmed/eliminated
4. Solve - Fix root cause, verify fix, add regression test
5. Prevent - Analyze why it happened, implement preventive measures

Features

Common Patterns

Copy# Define Phase
BOUNDARY DEFINITION
What IS the problem:
- Specific error: [error message]
- Steps to reproduce: [steps]
- When started: [date/commit]

What is NOT the problem:
- What still works: [list]
- Red herrings eliminated: [list]

Success criteria:
- What does "solved" look like?

# Hypothesize Phase
H1: [Most likely] - Evidence: ___ - Test cost: Low
H2: [Second likely] - Evidence: ___ - Test cost: Medium
H3: [Less likely] - Evidence: ___ - Test cost: High

RULE: Generate hypotheses BEFORE testing any.

# Test Phase
Test H1: [15 min box]
  Action: ___
  Result: CONFIRMED / ELIMINATED / INCONCLUSIVE

Copy# Risk Assessment Matrix
| Hypothesis | Probability | Impact | Test Effort | Priority |
|------------|-------------|--------|-------------|----------|
| H1         | High        | High   | Low         | 1st      |
| H2         | Medium      | High   | Medium      | 2nd      |

Use Cases

• Debugging complex bugs when conventional approaches fail
• Troubleshooting production incidents with unknown root causes
• Analyzing intermittent failures or race conditions
• Investigating performance regressions
• Resolving integration issues between systems

Best Practices

Related Skills

See also these related skill documents for complementary techniques:

• debugging-systematically - Four-phase debugging framework
• tracing-root-causes - Deep investigation techniques
• thinking-sequentially - Structure reasoning as numbered thoughts
• verifying-before-completion - Ensure fix is complete