Build decision matrices with AI
Compare options using agreed criteria, evidence and explicit trade-offs
Build decision matrices with AI to make the basis of a choice visible. The decision-matrix skill establishes the decision, alternatives, deadline and decision owner, then organises value, cost and risk criteria. Weights are agreed before scoring and should total one hundred per cent rather than being adjusted to favour a preferred answer.
Each score needs evidence or a stated rationale. Weighted totals inform judgement; they do not decide automatically. Close results can prompt a comparison of forces supporting or resisting change. The recommendation records trade-offs, reversibility and dissent. Recalculate the numbers yourself: one worked total in the source contains a typo, although its example winner remains unchanged.
What it does
- Agree criteria before scoring alternatives
- Document evidence behind each score
- Calculate and compare weighted totals
- Explain the main trade-offs
- Assess reversibility and decision timing
- Preserve dissent alongside the recommendation
Install and use
- Load the folder and licence using the documented Claude Code route. Prepare the options, constraints and evidence; the host processes those inputs, with no separate scoring service required.
- For example: “Compare three inventory tools on total cost, integration and setup time. Agree weights with me before scoring.”
- Use a consistent scale and clarify what a higher score means. Review the weights, justify scores and independently recalculate totals; investigate close results rather than overstating small differences.
- Review the recommendation and what would change it. The actual decision owner approves any choice; producing the matrix does not place an order or make a commitment.
FAQ
Source
Made by qa-aman. Original skill: decision-matrix, published on GitHub under the MIT licence.
View on GitHub