01Executive Summary

The rapid proliferation of autonomous AI agents operating in blockchain environments has created an unaddressed systemic risk. These agents — executing trades, managing liquidity, processing payments, and handling sensitive data on behalf of human principals — collectively control billions of dollars in assets across leading high-performance chains. Yet no dedicated insurance infrastructure covers them. When agents fail, are exploited, or breach data obligations, their operators bear the entire loss with no mechanism for recovery.

Regore is a decentralised risk-sharing protocol that fills this gap. By combining parametric trigger logic, battle-tested oracle infrastructure, and a transparent on-chain reserve pool, Regore delivers the first AI-agent-native insurance products designed for high-throughput blockchain ecosystems. Three coverage lines address the most material risks operators face:

• Financial Loss Coverage: up to $30,000 per session, triggered automatically when wallet losses exceed a defined threshold.
• Per-Trade Cover: up to $5,000 per transaction, settling on the price delta between order submission and execution.
• Data Breach Indemnity: up to $30,000 per incident, resolved by a five-member staked reviewer panel within 72 hours.

Regore operates as a decentralised autonomous mutual — modelled on the legal and operational structure validated by Nexus Mutual — requiring no insurance licence and governed by its community of token holders. Premiums are priced algorithmically based on each agent's on-chain risk profile. Reserve capital is pooled, deployed into low-risk DeFi yield strategies, and backstopped by a reinsurance arrangement for large individual claims. Liquidity providers earn risk-adjusted returns of 8–14% APY.

Regore is raising a seed round to fund smart-contract development, security auditing, oracle integration, and the onboarding of design-partner operators for a controlled mainnet launch.

02The Problem: Uninsured AI Agents in a High-Value Environment

2.1The Rise of Autonomous AI Agents On-Chain

The combination of sub-second finality, near-zero transaction costs, and a rich ecosystem of DeFi primitives across leading high-performance chains has made these networks the preferred environment for high-frequency, programmatic activity. AI agents — software systems that autonomously execute actions in pursuit of defined objectives — have become significant participants. They arbitrage price discrepancies across decentralised exchanges, rebalance lending positions, process micro-payments on behalf of users, and increasingly manage sensitive personal and financial data.

By early 2026, total value locked across major DeFi ecosystems exceeded tens of billions of dollars. Industry analysts estimate that autonomous agents — including algorithmic trading bots, AI-powered portfolio managers, and automated payment rails — account for a growing share of on-chain transaction volume. The global market for AI-agent transaction infrastructure is projected to reach $40 billion by the end of 2026.

2.2The Insurance Gap

Despite their scale and sophistication, AI agents are categorically uninsured. Three data points illustrate the severity of this gap:

Existing DeFi insurance protocols — including Nexus Mutual, InsurAce, and Unslashed Finance — are structurally mismatched for AI-agent use cases. They are designed primarily for Ethereum smart-contract risk, require manual claim filing and adjudication, and offer no per-transaction coverage mechanism. None are deployed on Solana. Armilla AI has pioneered enterprise AI liability coverage, but their product targets corporate clients via traditional insurance channels, not on-chain agent operators that require real-time, automated settlement.

2.3Categories of AI Agent Risk

Through research and operator interviews, Regore has identified three primary categories of loss event that AI agent operators consistently cite as material concerns:

Financial Loss

An agent's wallet is drained through exploitation of a smart-contract vulnerability, a flash-loan attack on a protocol it is interacting with, or a configuration error that causes runaway adverse trading. The operator loses the capital under management with no recourse. This category is particularly severe because agents often operate with positions that significantly exceed the value of the underlying protocol's treasury.

Per-Trade Slippage and Execution Failure

High-frequency agents are systematically exposed to the delta between the price at which an order is submitted and the price at which it executes — a gap caused by block-time latency, network congestion, or MEV extraction. While individual slippage events may be small, they compound across thousands of trades per session and can render otherwise profitable strategies uneconomic. No existing product insures individual transaction execution.

Data Breach

AI agents increasingly handle sensitive data: user wallet addresses, trading strategies, personally identifiable information, and proprietary model weights. A breach — whether through a compromised API key, an insecure storage mechanism, or a man-in-the-middle attack on an agent's communication channel — can expose operators to regulatory liability (particularly under emerging AI-accountability frameworks) and reputational damage that existing indemnity products do not cover in the decentralised context.

"When an agent is hacked or leaks data, the operator loses everything. There is no call centre, no adjuster, no policy. Just loss."

This absence of a coverage layer is not merely a commercial inconvenience: it constitutes a systemic vulnerability that suppresses the legitimate growth of AI-agent infrastructure. Operators who would otherwise deploy capital at scale are constrained by tail-risk exposure they cannot hedge.

03The Regore Protocol

3.1Design Philosophy

Regore is built on four first principles derived from the failure modes of existing DeFi insurance products:

• Parametric over indemnity: wherever possible, payouts trigger from objectively measurable on-chain events rather than subjective loss assessment. This eliminates adjuster fraud, moral hazard from inflated claims, and the delays inherent in manual review.
• Oracle-native: coverage triggers are derived from the same oracle infrastructure that the broader DeFi ecosystem relies upon, ensuring that the definition of a loss event is tamper-resistant and verifiable by all parties.
• Solana-native: all premium payments, reserve management, claim payouts, and governance actions are settled on Solana using USDC. No off-chain intermediaries, no custodians, no trust assumptions.
• Risk-aligned incentives: every participant in the system — policyholders, liquidity providers, and reviewers — must have economic skin in the game that aligns their behaviour with honest, sustainable operation of the protocol.

3.2Coverage Products

Product 1 · Financial Loss Coverage

Financial Loss Coverage protects AI agent operators against material wallet depletion events. At policy inception, the operator registers their agent's primary wallet address and selects a loss threshold expressed as a percentage of the insured balance (minimum 10%, maximum 50%). They lock a 10% co-insurance deposit into an escrow account.

A wallet-monitoring oracle observes the wallet balance at 60-second intervals throughout the policy period. When the reported balance falls below the threshold, an independent price feed is used to compute the USD-denominated loss at the moment of breach. The protocol's smart contract automatically transfers the lesser of the computed loss amount or the policy maximum ($30,000) in USDC to the operator's registered wallet. No claim filing is required.

Product 2 · Per-Trade Cover

Per-Trade Cover insures individual on-chain transactions against adverse price movement between order submission and execution. This product is designed for high-frequency trading agents, market-making bots, and arbitrage strategies where per-transaction slippage is a material operating cost.

When the agent submits a transaction, the Regore co-processor records the oracle-reported price of the relevant asset at submission time. After confirmation, the co-processor computes the execution price from the transaction's instruction data. If the negative delta — expressed as (submission price − execution price) / submission price — exceeds the operator's configured threshold, a payout is triggered. A small per-transaction micro-premium is deducted from the operator's pre-funded USDC deposit for each covered transaction, regardless of whether a payout occurs.

Product 3 · Data Breach Indemnity

Data Breach Indemnity covers losses arising from the unauthorised disclosure, exfiltration, or misuse of data handled by a registered AI agent. At policy inception, the operator submits a SHA-256 hash of their agent's execution logs for the prior 30 days — a requirement that establishes a cryptographic baseline of the agent's normal operational footprint and serves as the ground truth for breach verification.

When a data breach occurs, the operator submits a claim package comprising: the incident timestamp, a description of the breach mechanism, a signed assertion of loss amount, and execution logs sufficient to demonstrate the breach event. Five reviewers, selected randomly from Regore's staked reviewer registry, are allocated the claim. Reviewers earn a review fee if their vote aligns with the majority outcome; reviewers in the minority forfeit a portion of their stake. A 3-of-5 supermajority is required for claim approval. If approved, USDC up to the policy maximum ($30,000) transfers to the operator within 72 hours of claim submission.

04Protocol Architecture

4.1Smart Contract Stack

Regore is implemented as a set of programs deployed on Solana using the Anchor framework. The core programmes are:

• Registry Programme: maintains the on-chain registry of agent wallet addresses, coverage parameters, policy states, and oracle authorisation records.
• Premium Vault Programme: accepts USDC premium payments and co-insurance deposits, routes funds to the reserve pool, and manages escrow release on policy cancellation or claim settlement.
• Claim Processor Programme: receives oracle attestations for parametric products and executes automatic transfers; receives signed reviewer votes for data-breach claims and executes conditional transfers upon quorum.
• Reserve Pool Programme: manages the aggregate pool of LP capital and unearned premiums; authorises transfers to the claim processor; interfaces with reputable lending markets for idle-capital yield generation.
• Governance Programme: implements REGR token-weighted voting on protocol parameter changes, premium model updates, reviewer eligibility criteria, and reserve deployment decisions.

All programmes are formally verified against a specification that constrains state transitions to prevent reentrancy, integer overflow, and unauthorised fund access. The codebase will be audited by a reputable security firm prior to mainnet deployment.

4.2Oracle Architecture

Regore's oracle architecture is deliberately layered, with each layer designed to handle a specific category of information and provide a distinct set of security guarantees:

Layer 1 · Parametric Oracle

A pair of decentralised oracle networks provides real-time price feeds and configurable off-chain data feeds (including wallet balance polling and custom computation tasks). For Financial Loss Coverage and Per-Trade Cover, these feeds serve as the sole source of truth: if the data says a loss occurred, the payout executes. No human discretion is applied. This layer handles approximately 95% of expected claims by volume.

Layer 2 · Hybrid Staked Review (UMA-style Optimistic Oracle)

For Data Breach Indemnity — where the loss event is a real-world occurrence that cannot be directly observed on-chain — Regore implements a reviewer panel modelled on UMA Protocol's optimistic oracle and the Nexus Mutual claim-assessment framework. Reviewers must stake a minimum of 1,000 REGR tokens to join the registry. For each claim, five reviewers are selected using a verifiable random function (VRF) seeded by the claim transaction hash.

Reviewer incentives are structured to produce honest outcomes: the expected return from voting truthfully (earning the review fee plus stake preservation) exceeds the expected return from colluding with a fraudulent claimant (earning a share of the fraudulent payout minus stake forfeiture risk). This property holds as long as the proportion of dishonest reviewers remains below 40%, which the staking requirement and slashing mechanism are calibrated to achieve.

Layer 3 · DAO Governance Escalation

Any party — claimant, reviewer, or liquidity provider — may escalate a disputed data-breach claim to the protocol's DAO governance layer, implemented via the Governance Programme. Disputed claims are put to a token-weighted vote among all REGR holders, with a seven-day deliberation period. This layer serves as the final backstop for edge cases that the reviewer panel mechanism cannot definitively resolve.

4.3Pricing Model

Premium pricing is computed on-chain at policy inception using a base rate multiplied by four risk factors derived from the agent's on-chain history and declared operational parameters:

The Capital Multiplier scales with the ratio of capital under management to the operator's historical average, penalising sudden leverage increases. The Data Sensitivity Factor reflects the classification of data the agent handles (from 1.0× for public data to 2.5× for regulated PII). The Autonomy Level Factor reflects the degree of human oversight in the agent's decision loop (from 1.0× for supervised agents to 1.8× for fully autonomous agents). All factors are declared by the operator and verified against on-chain evidence where available; misrepresentation voids coverage.

05Risk Management Framework

5.1Reserve Architecture

The Regore reserve pool is structured to ensure solvency under realistic stress scenarios while providing competitive returns to liquidity providers. At launch, the protocol targets a coverage-to-reserve ratio of 1.5:1 — meaning aggregate coverage limits outstanding will not exceed 150% of the reserve pool's USDC balance. This ratio is enforced programmatically: new policy issuance is paused if the ratio would be breached.

A minimum surplus fund of 15% of total premiums is retained before any distributions are made to liquidity providers. This fund accumulates over time and provides a first-loss buffer ahead of the general reserve pool.

5.2Fraud Suppression Mechanisms

Regore's fraud suppression stack combines four independent mechanisms, each targeting a different attack vector:

• 30-day waiting period: prevents purchase of coverage after a known exploit event has occurred but before it is publicly disclosed. This eliminates the most common form of insurance fraud observed in DeFi protocols.
• 10% co-insurance deposit: requires policyholders to bear 10% of the first-loss exposure. Forfeited 100% on fraudulent claims and 50% on claims rejected for insufficient evidence. Creates real financial disincentive against speculative or inflated claims.
• 30-day proof-of-operation requirement (Data Breach): establishes that the agent was genuinely operational before the alleged breach, preventing the creation of phantom agents for the purpose of filing fabricated claims.
• 50% payout cap during the first 90 days: limits maximum exposure during the period before the protocol has accumulated sufficient historical loss data to validate its pricing model.

5.3Reinsurance Arrangement

Claims exceeding $20,000 on any individual policy are eligible for reinsurance recovery under an arrangement under negotiation with a leading decentralised reinsurance protocol. Reinsurance participation provides a capital backstop that allows Regore to offer coverage limits that would otherwise be imprudent relative to the reserve pool size at launch. The reinsurance premium is treated as a protocol operating expense, borne by the protocol surplus fund rather than passed directly to policyholders.

5.4Reserve Deployment

Idle reserve capital not required to satisfy near-term claim obligations is deployed into vetted DeFi lending markets, generating a base yield of approximately 4–6% APY. This yield supplements the premium income distributed to liquidity providers, improving their effective return without increasing policyholder costs. The protocol governance defines a maximum deployment cap — initially 40% of the reserve pool — to ensure sufficient liquidity for immediate claim settlement.

06Tokeneconomics

6.1The REGR Token

REGR is the native governance and utility token of the Regore protocol. Its functions are:

• Governance: REGR holders vote on protocol parameters, premium model updates, reserve deployment strategies, reviewer eligibility criteria, and treasury allocations.
• Reviewer staking: Reviewers must stake a minimum of 1,000 REGR tokens to participate in Data Breach Indemnity claim assessment. Stakes are slashed for dishonest or negligent review.
• Protocol fee discounts: Operators holding REGR receive a discount of up to 15% on their monthly premium, creating demand-side utility that does not depend on speculative value.
• LP participation: Liquidity providers who stake REGR alongside their USDC deposits receive boosted yield allocations, incentivising long-term capital commitment.

REGR has a fixed total supply of 100,000,000 tokens. It is not launched at protocol inception; the token launch is scheduled for a future milestone, after the protocol has demonstrated sustainable premium revenue and a loss ratio within the modelled range. This sequencing avoids the governance capture and speculative dynamics that have undermined other DeFi insurance protocols that launched tokens before establishing product-market fit.

6.2Revenue Distribution

Gross premium revenue is distributed in the following waterfall:

6.3Liquidity Provider Economics

Liquidity providers deposit USDC into the Regore reserve pool. Their capital is at risk in the event of claims exceeding premium income and the surplus fund. In exchange, they receive a pro-rata share of net premium income after the protocol surplus allocation and operating expenses, a pro-rata share of yield generated by the deployed reserve capital, and REGR token boost allocations proportional to their staked REGR balance (post-token launch).

Under the base-case model (loss ratio: 60%, blended premium consistent with risk-adjusted pricing, reserve deployment yield: 5%), the projected LP return is 11.4% APY. Liquidity providers are subject to a seven-day unbonding period before withdrawal, which prevents liquidity runs in the immediate aftermath of a large claim event.

07Market and Competitive Landscape

7.1Addressable Market

Regore's addressable market sits at the intersection of three converging trends: the growth of AI-agent infrastructure, the expansion of DeFi on high-performance chains, and the emergence of regulatory frameworks that impose liability on AI operators. The TAM is estimated as follows:

7.2Competitive Analysis

No existing protocol directly competes with Regore's combined value proposition. The competitive landscape can be segmented into the following:

Regore's structural advantages are: (1) first-mover for any AI-agent insurance product on high-performance chains; (2) the only protocol offering per-trade parametric coverage; (3) native integration with the oracle infrastructure already trusted by the protocols Regore's policyholders interact with; and (4) a product design that explicitly optimises for the operational requirements of autonomous agents rather than adapting a human-centric insurance model.

08Traction and Validation

8.1Product Development

Regore has completed the following milestones prior to its seed raise:

• Full-stack frontend MVP deployed on testnet, including Web3 wallet integration, a three-product policy purchase flow, a live agent monitoring dashboard, a per-trade feed interface, an LP deposit and withdrawal interface, and a claims submission portal.
• Decentralised oracle feeds integrated on testnet. Parametric trigger logic for both Financial Loss Coverage and Per-Trade Cover has been tested against real price data, with sub-second latency confirmed for Per-Trade Cover and 60-second polling confirmed for Financial Loss Coverage.
• SHA-256 log hash commitment mechanism designed and prototyped. The reviewer selection and staking interface is designed and pending smart-contract implementation.

8.2Customer Discovery

Regore conducted structured interviews with ten AI-agent operators across three categories: DeFi trading bots (four), data-processing agents (three), and payment automation agents (three). Key findings:

• Eight of ten operators confirmed willingness to pay a meaningful monthly premium for Financial Loss Coverage on their production agents, representing a strong signal of product-market fit.
• Six of ten operators cited per-trade slippage as a 'significant' or 'material' operating cost, validating the economic case for Per-Trade Cover.
• All ten operators were unaware of any existing insurance product that covered their agents' operations, confirming the absence of incumbent competition.

8.3Ecosystem Timing

The market infrastructure for Regore's product is maturing rapidly. In March 2026, Aon — one of the world's largest insurance brokers — completed the first on-chain insurance premium settlement on Solana. Armilla AI secured Lloyd's of London backing for the first AI model liability policy in April 2025. These developments signal that institutional capital and traditional risk-transfer markets are actively seeking exposure to AI insurance risk — a validation of the category that Regore is designed to serve natively.

8.4Liquidity Provider Interest

Three DeFi treasury teams have engaged in preliminary conversations about participating as liquidity providers in the Regore reserve pool at the target return of 11.4% APY. One investment fund focused on on-chain risk infrastructure has expressed interest in committing capital at protocol launch, subject to completion of the security audit and mainnet deployment.

09Go-to-Market Strategy

9.1Phase 1 · Controlled Launch

The controlled launch phase prioritises quality over quantity. Regore will onboard a maximum of ten design-partner operators recruited from the customer-discovery process. These operators receive subsidised premiums in exchange for committing to provide detailed operational feedback, participate in the reviewer registry with their technical staff, and serve as public references for the protocol. The design-partner programme allows Regore to validate the claim-trigger logic, pricing model, and reviewer-incentive structure against real-world loss events before opening the protocol to permissionless access.

Target metrics at end of Phase 1: $500,000 USDC in reserve pool TVL; ten active policies; zero unresolved disputed claims; loss ratio within modelled range (55–65%).

9.2Phase 2 · Ecosystem Growth

Phase 2 opens the protocol to permissionless policy purchase and LP participation. The Data Breach Indemnity product is launched, and the reviewer registry is opened to any REGR staker meeting the minimum stake threshold. Regore pursues integrations with the major DeFi protocols whose users are most likely to be operating AI agents: DEX aggregators, MEV infrastructure, perpetuals venues, and NFT automation platforms. A B2B API is launched that allows AI-agent platforms and agent-building frameworks to embed Regore coverage as a native feature of their products.

Target metrics at end of Phase 2: $5,000,000 USDC in reserve pool TVL; 200 active policies; 50 active reviewer stakers; loss ratio within modelled range.

9.3Phase 3 · Multi-Chain Expansion and Token Launch

Phase 3 deepens Regore's Solana integration — expanding oracle coverage, adding Solana-native yield strategies for the reserve pool, and launching the REGR governance token. An enterprise licensing programme is introduced for companies deploying AI agents that require on-chain-verifiable insurance indemnity for regulatory compliance.

Target metrics at end of Phase 3: $50,000,000 USDC in reserve pool TVL; 2,000 active policies on Solana; REGR governance fully active.

10Financial Model

10.1Revenue Projections

The following projections are based on base-case assumptions: blended monthly premium consistent with risk-adjusted pricing, 60% loss ratio, and reserve deployment yield of 5% APY. These assumptions are conservative relative to comparable DeFi insurance protocols.

10.2Use of Seed Funds

The seed round will be allocated as follows:

11Team

11.1Current Team

Regore was founded by a team with deep expertise in AI-agent infrastructure, DeFi protocol design, and product strategy. The founding team has built and shipped full-stack Web3 products from inception to deployment, with experience spanning financial advisory, product management, and blockchain development.

11.2Open Roles

Regore is actively recruiting for two critical roles:

• Chief Technology Officer / Co-Founder: senior smart-contract developer with demonstrated expertise in DeFi protocol architecture across high-performance chains. This person will own the smart-contract architecture, oracle integration, security-audit process, and devnet-to-mainnet deployment. Equity stake available; compensation benchmarked to market rates for senior protocol engineers.
• Head of Risk: actuarial background or DeFi insurance protocol experience. Will design and maintain the pricing model, calibrate reserve ratios, structure the reinsurance programme, and build the loss-ratio model from historical DeFi exploit data. Reporting to the CEO.

Regore is seeking advisors from organisations including Nexus Mutual, leading reinsurance protocols, major oracle networks, Armilla AI, and prominent AI-agent protocols. Introductions to qualified candidates for open roles are welcomed.

12Legal Structure and Regulatory Considerations

12.1Decentralised Mutual Model

Regore is structured as a decentralised risk-sharing mutual, not as a licensed insurance company. This structure — pioneered by Nexus Mutual and validated through several years of operation and regulatory scrutiny — does not require the protocol to hold an insurance licence in most jurisdictions because coverage is provided through a discretionary mutual rather than a regulated insurance contract. Members of the mutual purchase cover from a shared pool of capital contributed by other members and LP participants.

This structure is legally distinct from insurance in several important respects: coverage is discretionary (the protocol's smart contracts determine outcomes algorithmically rather than through contractual obligation), membership is open to any person meeting the registration criteria, and the mutual is governed by its token-holding members rather than by corporate directors. Regore's legal counsel is reviewing the applicability of this structure under relevant jurisdictions, with a focus on Singapore, the British Virgin Islands, and the Cayman Islands as potential domiciles for the protocol's legal entity.

12.2Disclaimers

Regore does not offer licensed insurance products. Coverage provided by the Regore protocol is discretionary mutual coverage, not an insurance contract. Policyholders should be aware that the protocol's smart contracts, while subject to rigorous security auditing, may contain undiscovered vulnerabilities. Past performance of DeFi insurance protocols does not guarantee future performance of Regore. The REGR token is a utility and governance token and is not offered as a security or investment.

13Conclusion

The autonomous AI agent is the defining new participant in decentralised finance. These agents now control billions of dollars in assets, execute millions of transactions, and handle sensitive data at a scale and speed that no human operator can match. Yet they remain categorically uninsured — a gap that suppresses the legitimate deployment of capital, exposes operators to unhedgeable tail risk, and creates systemic fragility in an ecosystem that is otherwise rapidly maturing.

Regore addresses this gap with a precision-engineered, chain-native parametric insurance protocol. By combining decentralised oracle infrastructure with a rigorous staked-reviewer mechanism for non-parametric claims, Regore delivers coverage that settles in seconds for parametric events and within 72 hours for data-breach incidents — without brokers, adjusters, or paper forms.

The market opportunity is large, the competitive field is empty, and the technical infrastructure to build this product now exists on Solana. Regore is building the insurance layer that autonomous AI agents require and that the ecosystem urgently needs.

"No AI agent should operate without coverage. Regore is building the infrastructure to make that the default."