What makes Non-Zero-Sum notable?

1.1K downloads. Trust evidence available

How should Non-Zero-Sum be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Non-Zero-Sum

Activate when: someone says 'this is win-lose,' 'we can't both win,' 'what's in it for them to cooperate,' 'is there a deal here,' or 'how do we get past thi... Skill: Non-Zero-Sum Owner: deciqai Summary: Activate when: someone says 'this is win-lose,' 'we can't both win,' 'what's in it for them to cooperate,' 'is there a deal here,' or 'how do we get past thi... Tags: latest:1.0.5 Version history: v1.0.5 | 2026-07-16T18:08:51.281Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/non-zero-sum.json) v1.0.4 | 2026-07-09T11:19:17.648Z | user R

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.1k

Updated

Oct 11, 2026

Version

1.0.5

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.1K downloads reported by the source. Last updated 10/11/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 11, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 11, 2026
Adoption signal: 1.1K downloadsadoption · observed Oct 11, 2026
Latest release: 1.0.5release · observed Jul 16, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:non-zero-sum

Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-deciqai-non-zero-sum/snapshot"

Documentation

CLAWHUB

127,809 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: non-zero-sum
description: "Activate when: someone says 'this is win-lose,' 'we can't both win,' 'what's in it for them to cooperate,' 'is there a deal here,' or 'how do we get past this standoff'; a negotiation or conflict feels deadlocked; you're designing a platform, contract, or institution that needs to align competing parties.
Do NOT activate when: the resource pool is genuinely fixed and one-shot with no side effects (true zero-sum); the conflict is identity- or values-based with no concrete trade that creates net value. More: deciqai.com/c/non-zero-sum"
---

# Non-Zero-Sum

## Overview

A non-zero-sum interaction is one where mutual gain (or mutual loss) is possible — the parties' outcomes do not simply cancel each other out. Most real-world conflicts and negotiations are not zero-sum, but *feel* zero-sum because we focus on the visible resource rather than underlying interests. Robert Axelrod's computer tournament showed cooperation can emerge without central authority when interactions repeat and the future is valued. Robert Wright extended this: the arc of history is driven by accumulating non-zero-sum arrangements — specialization, trade, institutions.

**Compose with neighbors:** Use prisoners-dilemma to model the payoff structure first. Use repeated-games-reputation when the key variable is whether interaction repeats. Use nash-equilibrium to find whether a stable cooperative outcome exists.

## When to Use

- A negotiation or conflict is deadlocked in zero-sum framing — each side treating every gain as the other's loss
- You want to find latent cooperative value in an adversarial relationship
- Designing an institution, platform, or contract to align incentives for competing parties
- Someone says: *"this is win-lose," "we can't both win," "what's in it for them," "could we cooperate instead of compete?"*
- A market is framed as winner-take-all — *"AI will take all the jobs / margin," "the AI capex will only pay off for the platform," "AI-native startups will crush incumbents (or vice versa)"* — and you need to test whether the layers can grow together instead

**When NOT to use:**
- Genuinely fixed-pool, one-shot interaction with no side effects — non-zero-sum framing is wishful, not analytical
- Interests are fundamentally incompatible (ideological, identity-based) with no concrete trade creating net value
- The real constraint is power asymmetry → use batna-zopa instead

## Coaching Novices (Adaptive Front Door)

- **Engine mode:** user has a specific standoff or competitive dynamic → run The Process directly.
- **Coach mode:** user is unfamiliar or has no concrete case → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

1. One-line what-it-is: non-zero-sum means both parties can gain — or both can lose — from an interaction. Most conflicts feel zero-sum but aren't.
2. Check fit against When to Use / When NOT to use — if genuinely

_meta.json

{
  "ownerId": "kn754b8sk22s8c6gjxt02bftbn88q7ye",
  "slug": "non-zero-sum",
  "version": "1.0.5",
  "publishedAt": 1784225331281
}

references/sources.md

# Sources — non-zero-sum

> *Primary sources for the [non-zero-sum](../SKILL.md) skill.*

- **Axelrod, Robert.** *The Evolution of Cooperation.* Basic Books, 1984. **Primary source for the tournament, Tit-for-Tat results, and the conditions for cooperation.** Verbatim Overview quote from ch. 1. https://www.basicbooks.com/titles/robert-axelrod/the-evolution-of-cooperation/9780465021215/
- **Axelrod, Robert.** "Effective Choice in the Prisoner's Dilemma." *Journal of Conflict Resolution* 24, no. 1 (1980): 3–25. **First academic publication of the tournament results.** https://doi.org/10.1177/002200278002400101
- **Wright, Robert.** *Nonzero: The Logic of Human Destiny.* Pantheon, 2000. **Primary source for the historical-accumulation thesis.** Verbatim Overview quote from the Introduction. The book extends Axelrod's framework across human history, from hunter-gatherer trade to the emergence of nation-states.
- **Rapoport, Anatol, and Chammah, Albert M.** *Prisoner's Dilemma: A Study in Conflict and Cooperation.* University of Michigan Press, 1965. **Primary source for the Prisoner's Dilemma framework** on which Axelrod's tournament is built. Documents the payoff structure and early empirical work on cooperation.
- **Nowak, Martin A., and May, Robert M.** "Evolutionary Games and Spatial Chaos." *Nature* 359 (1992): 826–829. https://doi.org/10.1038/359826a0 — Extends Axelrod's results to spatial structure, showing that cooperation can emerge and persist even when reputation and direct reciprocity are absent, through spatial clustering of cooperators.

- **Microsoft.** "Microsoft and OpenAI extend partnership." Official Microsoft blog, January 2023 (building on the partnership first announced in 2019). **Contemporary source for the AI-ecosystem example** — documents the multi-year, repeated cloud-plus-model partnership structure that gives the interaction a strong shadow of the future. https://blogs.microsoft.com/blog/2023/01/23/microsoftandopenaiextendpartnership/
- **Falling AI inference cost and rising capability, 2024–2025.** Widely reported industry trend documented across successive OpenAI, Anthropic, and Google model and pricing announcements (see each provider's model release notes and API pricing pages). **Contemporary source for the "variable-pie" claim** in the AI-ecosystem example — lower cost-per-token expanded the set of profitable AI applications rather than merely redividing a fixed pool. Exact figures vary by model and date and are omitted where not verifiable as of early 2026.

**Not cited and why:** Popular descriptions of "win-win thinking" as a personality trait or negotiation attitude (as in Stephen Covey's *7 Habits of Highly Effective People*) are not cited here — this skill grounds non-zero-sum analysis in payoff-structure mathematics and evolutionary stability, not in motivational or attitudinal frameworks. Covey's framing is inspirational; Axelrod's is analytical.

examples/ai-ecosystem-value-creation-2024-2026.md

# Method in Action: The AI Ecosystem — Positive-Sum vs. "AI Eats Everything" (2024–2026)

> *Example for the [non-zero-sum](../SKILL.md) skill.*

The dominant popular framing of the 2024–2026 AI boom is zero-sum: "AI will take all the jobs," "the foundation-model labs will capture all the margin," "the cloud providers own everything, so app builders are just renting a commodity that will crush them." This example runs the skill's process over the three main layers of the AI stack — foundation-model providers (e.g., OpenAI, Anthropic, Google DeepMind), cloud/compute providers (e.g., Microsoft Azure, AWS, Google Cloud, plus the chip supplier Nvidia), and application builders (the startups and incumbents building products on top of the models) — to test whether the interaction is actually zero-sum, or whether the layers can grow together.

**Step 1 — Map positions vs. underlying interests:**

| Party | Stated position ("who wins the fixed pie") | Underlying interest |
| Foundation-model labs | "We own the intelligence layer; everyone above us is a thin wrapper." | Maximize inference demand and paying usage; need distribution to reach end users and real-world feedback to improve models. |
| Cloud / compute providers | "Compute is the bottleneck; we capture the rent." | Maximize durable, high-utilization compute consumption; need a growing population of workloads to justify enormous capex. |
| App builders | "We'll be commoditized the moment the labs move up-stack." | Ship differentiated products with proprietary data, workflow, and distribution; need cheaper, better, more reliable models. |

The *positions* are stated as a fight over one fixed pie of AI margin. The *interests* are not symmetric claims on one pool — each layer's interest is served by the *others' growth*: labs need distribution (app builders) and compute (cloud); cloud needs workloads (both); app builders need cheaper, better models (labs) running on reliable infrastructure (cloud).

**Step 2 — Construct the payoff matrix:**

Reduce to two representative players — the model/compute *platform* layer and the *app-builder* layer — and ask whether total value is fixed or variable.

| | App builders invest / build on the platform | App builders retreat / hedge away |
| **Platform invests in capability + access** | Both gain: cheaper, more capable models expand what apps can profitably do → more usage → more inference revenue and compute consumption. New categories (coding assistants, customer support, document workflows) become viable that did not exist before. | Platform loses: capex is stranded, utilization drops. Builders lose the capability they were waiting for. |
| **Platform extracts / locks in aggressively** | Builders lose margin and autonomy; platform gains short-term rent but starves the ecosystem that generates demand and feedback. | Both lose: builders route around the platform (open-weight models, multi-vendor abstraction), platform's demand shrinks, the category's growth sta

examples/axelrods-computer-tournament-1980.md

# Method in Action: Axelrod's Computer Tournament (1980)

> *Example for the [non-zero-sum](../SKILL.md) skill.*

Primary-source-documented case. Axelrod ran the first computer Prisoner's Dilemma tournament in 1980, inviting submissions from specialists in game theory across multiple disciplines. He ran a second tournament in 1981 with 62 entries. Both are fully documented in *The Evolution of Cooperation* (1984).

**Step 1 — Positions vs. interests:**
In the Prisoner's Dilemma (as in most real strategic interactions), the *position* of each player is "defect if the other defects, cooperate only if cooperation is guaranteed." The *interest* of each player is to maximize total accumulated payoff over repeated interactions. These are not the same.

**Step 2 — Payoff matrix:**
Mutual cooperation: both players receive R (reward). Mutual defection: both receive P (punishment, worse than R). Temptation: one defects while the other cooperates — defector receives T (temptation, highest single-round payoff); cooperator receives S (sucker's payoff, lowest). The standard ordering: T > R > P > S. The non-zero-sum gap: if both cooperate repeatedly (R per round) vs. if both defect repeatedly (P per round), the cooperation stream produces more total value. The gap is the non-zero-sum dividend.

**Step 3 — Shadow of the future:**
The tournament was designed as iterated (indefinitely repeated) interactions. Axelrod showed mathematically that the minimum condition for Tit-for-Tat to be the stable strategy is: the discount factor (how much each player values future payoffs) must exceed (T − R) / (T − P). When interactions are repeated and future payoffs are valued, defection becomes irrational even for purely self-interested players.

**Step 4 — Cooperation mechanism:**
Direct reciprocity (Tit-for-Tat) was the winning mechanism in both tournaments. Tit-for-Tat's structural properties: (a) it starts cooperating (not exploitable through pre-emptive defection); (b) it retaliates immediately (defectors receive P immediately, not with delay); (c) it forgives after one retaliation (the relationship can recover, unlike strategies that defect forever after one betrayal). These properties map precisely to what makes cooperation self-sustaining.

**Step 5 — First move:**
Axelrod's result: the optimal first move is unconditional cooperation, combined with credible and immediate retaliation capability. Strategies that started defecting never recovered: they generated defection spirals that left both parties worse off than the mutual cooperation equilibrium.

**What the tournament shows:** Cooperation can emerge and be sustained in non-zero-sum repeated interactions *without* any central authority, enforcement mechanism, or moral instruction. The self-sustaining mechanism is the payoff structure itself, combined with a sufficiently strong shadow of the future. The implication for institutional design: if you want cooperation, ensure the interaction is repeated, make defection i

Github ReposUpdated 2d agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.