Tanaka

Project Owner

AgentArena: Adversarial Agent Dev Platform

Expert Rating

n/a

Type SingularityNET RFP
Funding Request $50,000 USD
RFP Guidelines An AI-native Development Environment for the ASI:Chain

Overview

Ethereum's MEV crisis proved what happens when agents hit production untested against adversaries: billions lost. ASI:Chain's agent economy will face the same risk at larger scale. AgentArena is a competitive arena on DevNet where developers build agents and pit them against each other in economic scenarios — market making, governance, resource allocation — while an AI judge on Singularity Compute analyzes strategies, detects exploits, and recommends improvements. The arena IS the dev environment: agents evolve through adversarial pressure, then deploy to MainNet with zero code changes.

RFP Guidelines

Internal Proposal Review

An AI-native Development Environment for...

Ended on:

13 Feb. 2026

Days

Hours

Minutes

Type SingularityNET RFP
Total RFP Funding $50,000 USD
Proposals 27
Awarded Projects n/a

SingularityNET

Feb. 4, 2026

View RFP

This RFP seeks proposals for the development of an AI-native Development Environment (IDE) that improves the efficiency and accessibility of blockchain application development for the ASI:Chain.

Proposal Description

Our Team

Kyoto Digital fields a team with rare expertise at the intersection of multi-agent systems, blockchain development, and competitive AI. We have built agent competition platforms, deployed autonomous trading systems, and contributed to decentralized governance research. Our combined 30 years of experience span game-theoretic AI, Rholang development, and developer platform engineering — the exact skillset AgentArena demands.

Company Name (if applicable)

Kyoto Digital

Project details

Ethereum's MEV problem cost billions because agents were never stress-tested against adversaries before deployment. ASI:Chain faces the same risk at a larger scale — agents here will compete for resources, governance power, and economic influence, not just transaction ordering. Yet today, developers write agent logic, deploy to DevNet, and hope.

AgentArena fixes this by making competition the development environment.

HOW IT WORKS

Developers submit agents as Rholang processes or MeTTa programs. The Arena Engine deploys them to ASI:Chain DevNet — real BlockDAG consensus, real CBC Casper finality, not a simulator — and runs them in competitive economic scenarios:

• Market Making — agents provide liquidity under volatility shocks and adversarial arbitrage

• Governance — strategic voting under manipulation and coalition dynamics

• Resource Allocation — bidding for Singularity Compute capacity

• Escrow — multi-party negotiation with dispute resolution and collusion

After each round, an AI Judge running on Singularity Compute analyzes what happened: classifying strategies, detecting exploits with exact mechanisms, flagging emergent phenomena like price manipulation rings, and generating code-level recommendations referencing MeTTa type specs.

Developers then replay every decision their agent made — what it saw, what it chose, what happened next — and iterate. The cycle is: specify → compete → analyze → improve.

WHY THIS MATTERS

Unit tests cannot prepare agents for adversarial economics. You cannot write a test for "does my agent survive an attacker I haven't imagined." Emergent behavior only appears when populations interact under real incentives. AgentArena provides that pressure — and because it runs on actual DevNet, agents deploy to MainNet with zero code changes.

MeTTa's type system enables unique capabilities: type-checked strategies (proving an agent never bids more than its balance) and cross-agent analysis at the type level (the judge can formally reason about how two agents' typed interfaces interact).

Target: 80 active developers and 200+ unique agents across 4 arena types within 6 months.

Open Source Licensing

MIT - Massachusetts Institute of Technology License

Background & Experience

Our team lead designed a multi-agent simulation platform used by 3 blockchain protocols to test tokenomic models before MainNet launch. Our AI engineer built adversarial agent generators for DeFi security auditing, identifying $2.3M in potential exploits across 15 protocols. Our platform engineer developed competitive coding platforms serving 8,000+ developers with real-time match execution. We contributed to MeTTa agent specification discussions in the Hyperon community and authored a research paper on typed agent communication protocols at AAMAS 2024.

Links and references

https://foldspace.vercel.app/

Activity Summary

Milestones

3

Total

Discussion

0

Total Comments

Reviews

0

Total Posted

Project Team

2

Total People

Total Milestones
3
Total Budget
$50,000_USD
Last Updated
13 Feb 2026

Milestone 1 - M1: Arena Engine + DevNet Integration

Description

Build the Arena Engine: scenario definitions, agent deployment to DevNet as Rholang processes, real-time competition execution with BlockDAG consensus, scoring and ranking. Two arena types: Market Making and Escrow. MeTTa Agent Specification Format and spec-to-Rholang compiler. Agent lifecycle management.

Deliverables

1. Arena Engine managing agent deployment, scenario execution, result collection on DevNet. 2. Market Making Arena with configurable volatility and scoring. 3. Escrow Arena with dispute resolution and collusion scenarios. 4. MeTTa Agent Specification Format and spec-to-Rholang compiler. 5. Scoring/ranking system with persistent leaderboard. 6. Test suite: 20 scenario configurations.

Budget

$20,000 USD

Success Criterion

1. Manages 10+ simultaneous agents on DevNet without crashes across 50 runs. 2. Market Making top agent scores 30%+ above random baseline. 3. Escrow handles disputes correctly in 90%+ of 30 test cases. 4. MeTTa-to-Rholang compiler valid for 85%+ of 25-spec test suite. 5. All code open-sourced under MIT license.

Milestone 2 - M2: AI Judge + Adversarial Agent Gen

Description

AI Judge system: strategy classification, exploit detection, emergent behavior flagging, improvement recommendations. Adversarial agent generator creating AI opponents (front-running, manipulation, collusion). Both on Singularity Compute. Add Governance Arena and Resource Allocation Arena.

Deliverables

1. AI Judge: strategy classification, exploit detection, emergent behavior detection. 2. Improvement recommendations referencing MeTTa specs. 3. Adversarial agent generator with multiple attack strategies. 4. Governance Arena and Resource Allocation Arena. 5. Singularity Compute deployment for all AI workloads.

Budget

$18,000 USD

Success Criterion

1. Strategy classification 80%+ agreement with human experts. 2. Exploit detection finds 75%+ of injected exploits. 3. Adversarial agents defeat naive baselines in 80%+ of competitions. 4. Singularity Compute inference <8s per judge cycle (p95).

Milestone 3 - M3: Agent Studio + Replay + Beta Launch

Description

Web-based Agent Studio with MeTTa spec editor, type checking, Rholang compilation, one-click arena submission, step-by-step replay viewer. Beta launch with docs, 3 tutorial agents, leaderboard, community forum.

Deliverables

1. Agent Studio: MeTTa editor with type checking and Rholang panel. 2. Replay Viewer: step-by-step decisions with economic context. 3. 3 tutorial agents with guided exercises. 4. Persistent leaderboard and documentation site. 5. Public beta launch.

Budget

$12,000 USD

Success Criterion

1. New developers submit first agent within 45 minutes (4/5 testers succeed). 2. Strategy iteration cycle <5 minutes. 3. 25+ registered developers, 50+ agent submissions within 3 weeks. 4. User satisfaction 4.0+/5.0 from beta survey. 5. All code MIT licensed.

Join the Discussion (0)

Expert Ratings

Reviews & Ratings

No Reviews Avaliable

Check back later by refreshing the page.

The weighted average of the 4 perspectives Overall

0.0
Each RFP defines a maximum allowed budget, but teams can differentiate their proposal by offering a solution with a lower budget or a wider scope.Value for money

0.0
This rating indicates compliance to 'Must haves' but also adaptation of 'Nice to haves' and Non-functional requirements defined in the RFP.Compliance with RFP requirements

0.0
RFPs will offer varying degrees of freedom. This rating indicates the quality of the team's specific solution ideas, the provided details, and the reviewer's confidence in the team's ability to execute.Solution details and team expertise

0.0

Review Headline

0 /50 chars

Review Summary

0 /5000 chars

The weighted average of the 4 perspectives Overall

0.0
Each RFP defines a maximum allowed budget, but teams can differentiate their proposal by offering a solution with a lower budget or a wider scope.Value for money

0.0
This rating indicates compliance to 'Must haves' but also adaptation of 'Nice to haves' and Non-functional requirements defined in the RFP.Compliance with RFP requirements

0.0
RFPs will offer varying degrees of freedom. This rating indicates the quality of the team's specific solution ideas, the provided details, and the reviewer's confidence in the team's ability to execute.Solution details and team expertise

0.0

Review Headline

0 /50 chars

0 /5000 chars

Warning: Adding final group rating for this project will prevent expert users from adding new or editing existing reviews

Reviews and Ratings in Deep Funding are structured in 4 categories. This will ensure that the reviewer takes all these perspectives into account in their assessment and it will make it easier to compare different projects on their strengths and weaknesses. Overall (Primary) This is an average of the 4 perspectives. At the start of this new process, we are assigning an equal weight to all categories, but over time we might change this and make some categories more important than others in the overall score. (This may even be done retroactively). Feasibility (secondary) This represents the user\'s assessment of whether the proposed project is theoretically possible and if it is deemed feasible. E.g. A proposal for nuclear fission might be theoretically possible, but it doesn’t look very feasible in the context of Deep Funding. Viability (secondary) This category is somewhat similar to Feasibility, but it interprets the feasibility against factors such as the size and experience of the team, the budget requested, and the estimated timelines. We could frame this as: “What is your level of confidence that this team will be able to complete this project and its milestones in a reasonable time, and successfully deploy it?” Examples:

A proposal that promises the development of a personal assistant that outperforms existing solutions might be feasible, but if there is no AI expertise in the team the viability rating might be low.
A proposal that promises a new Carbon Emission Compensation scheme might be technically feasible, but the viability could be estimated low due to challenges around market penetration and widespread adoption.

Desirability (secondary) Even if the project team succeeds in creating a product, there is the question of market fit. Is this a project that fulfills an actual need? Is there a lot of competition already? Are the USPs of the project sufficient to make a difference? Example:

Creating a translation service from, say Spanish to English might be possible, but it\'s questionable if such a service would be able to get a significant share of the market

Usefulness (secondary) This is a crucial category that aligns with the main goal of the Deep Funding program. The question to be asked here is: “To what extent will this proposal help to grow the Decentralized AI Platform?” For proposals that develop or utilize an AI service on the platform, the question could be “How many API calls do we expect it to generate” (and how important / high-valued are these calls?). For a marketing proposal, the question could be “How large and well-aligned is the target audience?” Another question is related to how the budget is spent. Are the funds mainly used for value creation for the platform or on other things? Examples:

A metaverse project that spends 95% of its budget on the development of the game and only 5 % on the development of an AI service for the platform might expect a low ‘usefulness’ rating here.

A marketing proposal that creates t-shirts for a local high school, would get a lower ‘usefulness’ rating than a marketing proposal that has a viable plan for targeting highly esteemed universities in a scaleable way.
An AI service that is fully dedicated to a single product, does not take advantage of the purpose of the platform. When the same service would be offered and useful for other parties, this should increase the ‘usefulness’ rating.

Tanaka

Project Owner

Architects the Arena Engine and DevNet integration. Designs competition lifecycle, scoring, and agent deployment. Owns BlockDAG consensus and CBC Casper transaction ordering.

View Profile

Account Pending

Full Stack Dev

Develops Agent Studio, replay viewer, and leaderboard. Builds MeTTa spec editor with type checking and Rholang implementation panel. Owns web platform and beta launch.

Account Pending