Decision Matrix

Use this skill when the user faces a complex choice with multiple options and wants a systematic, quantified framework to evaluate trade-offs, test assumptions, and reach a defensible decision.

Good triggers

• "I need to choose between two job offers — help me decide."
• "Which car should I buy? Evaluate options quantitatively."
• "Decide between renting and buying a home."
• "Compare software vendors for my team."
• "Should I relocate to City A or City B?"
• "Help me make a structured decision about my career path."

Workflow

1. List options. Ask the user for 3-5 alternatives. If more than 5, suggest pre-filtering to the top contenders.

2. Define criteria. Ask the user to list the dimensions that matter for this decision. For each criterion, also capture which direction is better (higher = better, or lower = better). Examples: salary, commute time, growth potential, work-life balance, cost.

3. Weight criteria (1-10). Ask the user to assign importance weight to each criterion. Normalize to sum-to-100 percentages for clarity. Optional: flag criteria weighted > 9 as potential dealbreakers.

4. Score each option per criterion (1-10). Ask the user to rate each option against each criterion. If a criterion has objective data, suggest the score (e.g., salary in RMB scaled to 1-10).

5. Compute weighted scores. Calculate:

   WeightedScore(option) = Σ( score(option, criterion_i) × weight_i )
   

   Display as a heatmap:

   Option	Criterion 1 (w=30%)	Criterion 2 (w=20%)	...	Total
   A	8	6		7.2
   B	5	9		6.8

6. Sensitivity analysis. For each criterion, vary the weight by ±20% and re-rank:

   • If the #1 option changes under any ±20% weight shift, flag as unstable
   • Report the "stress test" — which criteria cause the most rank volatility
   • Identify the tipping point: what weight change would flip rank 1 and 2

7. Identify key differentiators. Find the criterion or criteria that most drive the ranking difference between the top 2 options. These are the "swing factors" the user should examine most carefully.

8. Deliver decision report. Structure:

   • Summary — top-ranked option with total score
   • Heatmap table — original scores and weighted totals
   • Sensitivity analysis — stability flag, tipping point if any
   • Key differentiators — the 1-2 criteria separating the leaders
   • Recommendation — ranked list with rationale per option
   • Caveats — assumptions made, data gaps, subjective scores flagged

Sample prompt

decision-matrix evaluate --options "Offer A,Offer B,Offer C" --criteria "薪资:8,发展空间:9,稳定性:6"