Thinking and Decisions skills
Thinking and Decisions Skills-md for AI Agents
Thinking and Decisions skills are SKILL.md files that help Claude, Codex and other AI agents reason through a choice instead of jumping to a generic answer. They cover framing the real question, picking the right decision framework for the stakes, stress-testing a plan with a premortem, and comparing options on a weighted trade-off so the decision, and the reason for it, is written down.
3 of 5 planned skills published, version 1.0.0, October 2026.
By Afeez Adebayo, Marketing, SEO & Growth Expert with 10+ years of experience.
Created 24 October 2026. Updated 24 October 2026. Free to use under CC BY 4.0.
Where to start: decision-frameworks to run the whole decision, then premortem-analysis and trade-off-analysis as the two methods it calls.
Decision-Making Frameworks
decision-frameworks · making decisions · 3,200 words · start here
Frames the real question, classifies a decision by reversibility and stakes, picks the lightest method that fits, and records the call with an owner and a revisit date.
Premortem and Decision Risk
premortem-analysis · making decisions · 2,651 words
Imagines a plan has already failed, works backwards to the causes, triages them into real threats and distractions, and sets owners, early-warning signs and measurable kill criteria.
Trade-off Analysis
trade-off-analysis · making decisions · 2,411 words
Compares options with a weighted decision matrix that avoids false precision: must-have gates, capped criteria, anchored whole-number scores and a sensitivity check that calls close results ties.
No skill matches that search. Clear it or choose All.
What these skills do
Most people make two mistakes with decisions. They agonise over small, reversible calls, and they rush the big, irreversible ones. They also judge past decisions by how they turned out rather than by whether the reasoning was sound. These skills fix both, by matching the effort to the stakes and writing the reasoning down so it can be reviewed.
Each skill is a self-contained SKILL.md. An agent reads the short description first, then loads the full file only when a task needs it.
Where to start
Start with Decision-Making Frameworks. It runs the whole decision and decides which method the choice deserves. It then calls Premortem and Decision Risk to stress-test a one-way door, and Trade-off Analysis to compare options on a weighted scorecard.
The skills and the sources behind them
Three skills in this set are live, with two more planned (problem framing and structured brainstorming).
| Skill | What it does | Grounded in |
|---|---|---|
| Decision-Making Frameworks | Frames the question, classifies by reversibility and stakes, picks the method, records the call | Bezos one-way and two-way doors; Bain RAPID; Atlassian DACI; Kahneman outside view |
| Premortem and Decision Risk | Imagines the plan has failed, surfaces the causes, sets kill criteria and owners | Gary Klein, HBR 2007; Mitchell, Russo and Pennington 1989; Shreyas Doshi risk triage |
| Trade-off Analysis | Compares options on weighted criteria, with must-have gates and a sensitivity check | Pugh concept selection; Saaty AHP; documented false-precision failure modes |
Example prompts
Each skill triggers on everyday wording:
- Help me decide between these two offers, and tell me how much process it needs.
- We are about to commit to this plan. What could make it fail?
- Compare these four vendors against what actually matters to us.
- Who owns this call, and how do we stop reopening it?
What the agent needs from you
These skills work best with a few inputs: the decision in your own words, the options so far, who makes the final call, the deadline, what a good outcome looks like, and how costly the choice would be to reverse. If something is missing, the skill asks for it rather than guessing, because ownership, deadline and reversibility change the whole approach.
What these skills will not do
They structure and record a decision. They do not make the decision for you, and they stop before any irreversible action, such as sending, paying, hiring or deleting, to get explicit human sign-off. They do not give legal, medical or financial advice; for decisions with those consequences they say plainly that a qualified professional should review the call. They do not fabricate numbers: every figure carries a source and a date, or is labelled an estimate and given as a range.
Frequently asked questions
What are Thinking and Decisions skills?
They are SKILL.md files that teach an AI agent to reason through a choice: frame the real question, match the effort to the stakes, stress-test a plan, compare options, and record the decision. They turn a vague request like help me decide into a structured, documented call.
Which skill should I install first?
Install Decision-Making Frameworks first. It runs the whole decision and decides which method the choice needs, then calls the premortem and trade-off analysis skills as required. The three work together, but Decision-Making Frameworks is the hub.
Do these skills replace my judgement?
No. They structure the thinking and record the reasoning, but a person makes and approves the decision. For any irreversible action, the skills stop and ask for explicit human sign-off rather than acting on your behalf.
Are they tied to one tool or product?
No. The methods are tool-agnostic and work in Claude, Cursor, Codex or any agent that reads SKILL.md files. Each skill names a tool-specific appendix separately from its core, so the thinking stays portable.
Are the frameworks evidence-based?
Yes. They draw on primary sources such as Gary Klein premortem research, Kahneman work on the outside view, Bezos reversible and irreversible decisions, and the decision-matrix literature. Every figure in the skills carries a source and a date.
Sources and more skills
Jeff Bezos, Amazon 2015 shareholder letter. Daniel Kahneman, Dan Lovallo and Olivier Sibony, “Before You Make That Big Decision,” HBR 2011. Gary Klein, “Performing a Project Premortem,” HBR 2007. Annie Duke, “Thinking in Bets” and “How to Decide.” Stuart Pugh, concept selection, 1981. Bain RAPID and Atlassian DACI.
See also the Productivity hub and the full skills-md library.
