problem-solver（问题解决者）

Problem Solver

A structured workflow for solving problems through iterative planning, reproduction, implementation, and validation.

When This Skill Activates

• User describes a bug or unexpected behavior
• User requests a feature implementation
• User mentions "fix", "debug", "implement", "solve", or "issue"
• Complex code changes requiring systematic approach

Workflow Overview

Phase 0: Input Collection (gather info from user)
    ↓
Phase 1: First-Pass Planning & Execution
    Step 1.1 → 1.2 → 1.3 → [1.4 ↔ 1.5 iterate until solved]
    ↓
Phase 2: Solution Validation
    ↓
    ├─→ Fundamental Fix → Complete
    └─→ Workaround → Return to Phase 1 with new approach

Phase 0: Input Collection

Purpose: Gather sufficient information from the user before analysis begins.

Required Information Checklist

Before starting, ensure you have:

• Scope: Files, directories, components involved
• Resources: Available tools, MCP servers, test commands
• Expected behavior: What should happen
• Actual behavior: What is happening instead
• Reproduction steps: How to trigger the issue (if known)

Clarification Protocol

If any required information is unclear or ambiguous:

1. Stop and ask - Do not assume or guess
2. Be specific - Ask targeted questions, not open-ended ones
3. Confirm understanding - Restate the problem before proceeding

Use AskUserQuestion tool when clarification is needed.

Phase 1: First-Pass Planning & Execution

Planning (Step 1.1-1.3) is the most critical part. Do not rush to implementation. A well-understood problem with a solid plan leads to a fundamental solution.

Step 1.1: Requirements Definition

Purpose: Structure and analyze the collected information into actionable requirements.

1. List all requirements as bullet points
2. Identify any ambiguous terms or conditions
3. Ask clarifying questions if needed
4. Get user confirmation before proceeding

Step 1.2: Information Gathering

1. Locate relevant code using Grep, Glob, Read
2. Trace data flow from input to output
3. Identify dependencies and side effects
4. Document findings for reference

Step 1.3: Reproduction Environment (Critical)

Establish a verifiable reproduction before making changes.

Choose appropriate method(s) based on the problem type. Combine as needed:

Methods are not mutually exclusive. For example, UI bugs may benefit from mock data + logging together, or a complex state bug might need all three approaches.

Verification checkpoint: Can you reliably trigger the issue?

Step 1.4: Implementation

Only proceed to implementation when Step 1.1-1.3 are thoroughly completed. Skipping planning leads to workarounds, not fundamental fixes.

1. Create task list with TodoWrite
2. Implement in small, testable increments
3. Run verification after each change
4. Document any deviations from plan

Step 1.5: Verification

• Issue no longer reproduces in test environment (run multiple times for intermittent issues)
• All existing tests pass
• Edge cases considered and handled
• No new warnings or errors introduced

For intermittent or timing-related issues, a single successful test is not sufficient. Run verification multiple times to ensure the fix is reliable.

Phase 2: Solution Validation

Analysis Questions

Answer each question honestly:

1. Root cause addressed? Does this fix the underlying cause, or mask symptoms?
2. Recurrence risk? Could similar inputs cause the same problem?
3. Side effects? Are other parts of the system affected?
4. Code quality? Is the solution clean and maintainable?

Decision Matrix

If Workaround Detected

1. Document why current approach is a workaround
2. Identify what a fundamental fix would require
3. Return to Phase 1 with new approach
4. Repeat until fundamental fix achieved

Completion Criteria

All must be true:

• Problem verified as resolved
• All tests passing
• Solution validated as fundamental (not workaround)
• Changes documented/committed appropriately

Quick Reference

For detailed checklists and examples, see:

• Analysis Checklist
• Reproduction Strategies