game-theorist

game-theorist

Model the other player's next move before you make yours.

Made by Thomas Sauvé at Salt & Silicon

---

Game theory for strategy, negotiation, pricing, fundraising, partnerships, and competitive decisions. Describe the situation. The skill turns it into a six-phase strategic model through dialogue.

Designed for humans using AI assistants. The assistant asks structured questions, maps incentives, tests likely responses, and ends with a concrete move.

It can miss context, overfit the framing you give it, or sound more confident than it should. Use it to pressure-test incentives and structure your thinking before you move. Please do not cheat on your soulmate because a markdown file told you to.

---

Install

Pick your agent. Run one command.

Agent	Install
Claude Code	claude plugin marketplace add saltandsilicon/game-theorist && claude plugin install game-theorist@game-theorist
Codex	npx skills add saltandsilicon/game-theorist -a codex
Cursor	npx skills add saltandsilicon/game-theorist -a cursor
Windsurf	npx skills add saltandsilicon/game-theorist -a windsurf
Cline	npx skills add saltandsilicon/game-theorist -a cline
Any other	npx skills add saltandsilicon/game-theorist

Install once. Invoke with /game-theorist or describe a strategic situation.

Manual install

git clone https://github.com/saltandsilicon/game-theorist.git
cp -r game-theorist/skills/game-theorist ~/.claude/skills/

For tools that load markdown instructions directly, point them at:

@./skills/game-theorist/SKILL.md

---

What is game theory?

Game theory is the study of decisions where other people can react.

It asks simple questions:

• Who are the players?
• What does each player want?
• What can each player do?
• What happens if everyone acts in their own interest?
• Which move changes the game in your favour?

You use it when advice like "just raise prices", "just negotiate harder", or "just wait" is too flat because someone else can answer your move.

---

What this skill does

Strategic decisions get hard when another player can react. This skill slows the analysis down: players first, incentives second, moves third, equilibrium fourth, recommendation last.

Six phases, one at a time:

1. Phase 1: Deconstruction, what game is being played? (often different from what you think)
2. Phase 2: Incentive Mapping, what do the players want, not what they say they want?
3. Phase 3: Strategy Space, what moves exist? which are dominated and can be eliminated?
4. Phase 4: Equilibrium, what outcome emerges if everyone acts rationally?
5. Phase 5: Recommendation, the concrete move, how to make it credible, and what to do if it fails
6. Phase 6: Dynamic Adaptation, reputation, tit-for-tat, and changing the rules of the game

---

A full conversation

You: I'm raising prices 20%. Main competitor has cheaper infrastructure. They've been matching my
     prices for two years. What happens?

Agent: Phase 1: Deconstruction.

       Sequential game, repeated. You are the Stackelberg leader, they react to your move. Two
       years of matching is a signal: "we will undercut if you diverge." That signal is the
       equilibrium you are about to break.

       Before I map incentives, one question. Their cost structure relative to yours:
       A) They are cheaper to operate
       B) Similar costs, they are matching strategically
       C) You don't know

You: A. Their unit economics are roughly 30% better.

Agent: Phase 2: Incentive Mapping.

       You: stated objective is margin recovery. True objective is testing whether the market will
       absorb a price increase without churn. BATNA: hold price, accept slower margin growth.
       Time horizon: long.

       Them: stated objective is "competitive pricing." True objective, given the cost asymmetry,
       is share capture. They have headroom to undercut you and stay profitable. BATNA: continue
       matching at current spread.