UniBrain Hook

Autonomous AI strategy execution for Uniswap v4 pools using Reactive Contracts. UniBrain turns pool telemetry into deterministic, authenticated, on-chain strategy updates without keepers or bot operators.

Problem

LP strategy operations in AMMs are still mostly manual or bot-driven:

• Strategy updates are delayed or inconsistent under volatile market conditions.
• Off-chain keepers create liveness and trust assumptions.
• Execution auth is often weak across origin and destination chains.
• Protocol teams cannot easily prove deterministic, replay-safe decisioning to judges, auditors, or partners.

Solution

UniBrain implements an event-driven strategy loop with strict on-chain verification:

• Origin: Uniswap v4 hook emits telemetry on pool activity.
• Reactive: deterministic AI logic extracts features, scores risk/regime, and computes bounded strategy decisions.
• Destination: executor authenticates callback source + RVM identity + nonce, then applies bounded updates to strategy state.

Integrations

• Uniswap v4 hook architecture (PoolManager, hook permission bits, swap hook flow)
• Reactive Network / Lasna (subscribe(...), react(LogRecord), callback payload path with RVM argument replacement)
• Base Sepolia as both origin and destination execution domain

Major Components

• contracts/src/UniBrainTelemetryHook.sol: telemetry emission and fee-target reads from strategy state
• contracts/src/UniBrainStrategyBook.sol: bounded strategy state, replay checks, fee/range storage
• contracts/src/UniBrainStrategyExecutor.sol: callback authenticity checks, RVM allowlist checks, non-reentrant execution
• reactive/src/UniBrainReactiveStrategy.sol: subscription lifecycle, deterministic decisioning, callback emission
• shared/solidity/src/DeterministicAI.sol: deterministic feature extraction and scoring model
• scripts/demo-sepolia.sh: end-to-end proof runner with detailed tx URL logging
• frontend/: dashboard for telemetry, decisions, and execution visibility

Diagrams and Flowcharts

1) User Perspective Flow

flowchart LR
  U["LP Operator"]
  D["Demo Script"]
  H["UniBrainTelemetryHook"]
  R["UniBrainReactiveStrategy"]
  E["UniBrainStrategyExecutor"]
  B["UniBrainStrategyBook"]

  U -->|Run demo script| D
  D -->|Generate pool activity swap| H
  H -->|Emit telemetry| R
  R -->|Compute decision and callback| E
  E -->|Apply decision| B
  B -->|Updated fee and range state| U

Loading

2) Architecture Flow (Subgraph View)

flowchart TB
  subgraph Origin[Origin - Base Sepolia]
    PM[PoolManager]
    HK[UniBrainTelemetryHook]
    PM --> HK
  end

  subgraph Reactive[Reactive - Lasna]
    RC[UniBrainReactiveStrategy]
    SYS[Reactive Service]
    SYS --> RC
  end

  subgraph Destination[Destination - Base Sepolia]
    EX[UniBrainStrategyExecutor]
    SB[UniBrainStrategyBook]
    EX --> SB
  end

  HK -->|Telemetry logs| RC
  RC -->|Callback payload| EX

Loading

3) AI Decision Lifecycle

sequenceDiagram
  participant PM as PoolManager
  participant H as UniBrainTelemetryHook
  participant R as UniBrainReactiveStrategy
  participant E as UniBrainStrategyExecutor
  participant B as UniBrainStrategyBook

  PM->>H: beforeSwap/afterSwap
  H-->>R: TelemetryEmitted(poolId,...)
  R->>R: deterministic feature extraction + scoring
  R-->>E: Callback(rvmId, decision payload)
  E->>E: proxy + RVM + nonce checks
  E->>B: applyDecision(...)
  B-->>E: applied=true

Loading

Live Deployments and TXID URLs (2026-03-07)

Current Addresses

• Owner: 0x4b992F2Fbf714C0fCBb23baC5130Ace48CaD00cd
• Base Sepolia PoolManager: 0x05E73354cFDd6745C338b50BcFDfA3Aa6fA03408
• Base Sepolia CallbackProxy: 0xa6eA49Ed671B8a4dfCDd34E36b7a75Ac79B8A5a6
• Base Sepolia StrategyBook: 0x25159820b8e4E1d7661bA50E0093b244fCCcCc22
• Base Sepolia StrategyExecutor: 0x7aD1473337C4E1b5B808e897b41B353b4cC5A597
• Base Sepolia UniBrainTelemetryHook: 0xB40207558471cA465978Dbb02e96E07d335b40C0
• Reactive Lasna UniBrainReactiveStrategy: 0x994987B24486717DDa19411c56d82c29f6B19CDC
• Reactive Lasna Service: 0x0000000000000000000000000000000000fffFfF

Deployment TX URLs

• StrategyBook create: https://sepolia.basescan.org/tx/0xb8f9963601051e1790ff250c377280ed140721d46ff57228fc3801f6a2c07f7b
• StrategyExecutor create: https://sepolia.basescan.org/tx/0xcb8a11603757f2d5779373638d54df0ba586f020ea8e62fa7fcf5497e2d7d114
• StrategyBook setExecutor: https://sepolia.basescan.org/tx/0x8bba29cea72a382bcbfd4288c84404dcf96b1c15ceb9ddffed8ff7a58e200b77
• Hook create2: https://sepolia.basescan.org/tx/0x9d4b7f372af50549847cb32a077fa5aff7bd57460aee02c3bfc150cb54dc0d4f
• Reactive deploy: https://lasna.reactscan.net/tx/0xadd29b5374baf40e9f21d498f1ec18368ac7f8b0a6b1d8e8f19867d4be402c3f
• activateSubscriptions: https://lasna.reactscan.net/tx/0x0862a271909fe75033035d9db6768c41b8afcdce457b8f0d2ce49c15838647c3

Demo Run Proof (latest: 2026-03-07T14:14:27Z)

Demo Script Validation Process

The make demo-sepolia flow validates:

1. Chain + bytecode preflight checks on Base Sepolia and Lasna.
2. Reactive health checks (subscriptionsActive, expected count, debt coverage).
3. Fresh Base Sepolia activity generation to guarantee telemetry.
4. Reactive-side proof via RNK RPC (DecisionComputed + Callback log detection).
5. Destination execution scan on Base (DecisionExecutionAttempt).

Proof TX URLs Printed by Demo

• Base telemetry reference tx: https://sepolia.basescan.org/tx/0x5709f71f9b27c6ed083d816265085ef812eeee4d114db7c96956287bb0df0539
• Lasna decision tx: https://lasna.reactscan.net/tx/0xc2051f9bd583a32dc966767a83331c6b921173e3e2c55dd56f612e800b2cf948
• Lasna callback tx: https://lasna.reactscan.net/tx/0xc2051f9bd583a32dc966767a83331c6b921173e3e2c55dd56f612e800b2cf948
• Activity poolId: 0xd3be67c93d778d581b23fb1396b7c7b86cd88b5c3047c9217ea52abf241607e1

Full Base Activity Batch (15 TX URLs)

• https://sepolia.basescan.org/tx/0xa832360ac56e7eb0ccbc96722c301af677e119e758da9ce30ba3eedaafd947ed
• https://sepolia.basescan.org/tx/0x0a75702549ebc107ba6e9768a24455d35e5f46220237a5e97a9e77e56aeecc8d
• https://sepolia.basescan.org/tx/0x0ccf6cbb90c0e6e729815e72c0a07c1531f756784d5b89f7b1293ba2a6a55c30
• https://sepolia.basescan.org/tx/0x7003da3dc30dd7453af7e506e63bf8aaf74dd763184cd6a9c691d9d1b43de41c
• https://sepolia.basescan.org/tx/0x5c9cb11088eb58c21e4752cf3b19b6743632217e0c716a732ab343005f7fd489
• https://sepolia.basescan.org/tx/0xa55ffb2fc29de868c87e1655576a2587a46f364b51b88d8dd70aa7633dac0c38
• https://sepolia.basescan.org/tx/0x3cc7d28a2ffa014ad87beece94bfe9e3ce1ba0a1a3deb5e1433469520c95bc44
• https://sepolia.basescan.org/tx/0x5a2f969a855b7e311714778d5481b2a0a483150cf150fcf2d62d8514403425ac
• https://sepolia.basescan.org/tx/0x226cceae329d5caf7d77e6d216c32c68d018ed3d53d102f9c6c8e941c0593c21
• https://sepolia.basescan.org/tx/0x9774e18a3863b46ff4d5e085341f11c9bcac7a7689bd72be3816e20590dce8c2
• https://sepolia.basescan.org/tx/0x85a5a6e86f34e5a98aedfd7849ed5abdf1d5a3f24ccbaaea2f28f28a8f77ec0e
• https://sepolia.basescan.org/tx/0x16e416d86950e7a611f0262d2ae3c065356c37b8f00fadcacfeace7c7c8870ad
• https://sepolia.basescan.org/tx/0xf40e1cd31cdc1cca2bc1d0f23315f30a4229b9af7bb731abf04de64b0460f7de
• https://sepolia.basescan.org/tx/0xaeb0ac67c74bbdfcc9d2080ab7a468dea53b998432c0fcf852a8aa7e1d19d5de
• https://sepolia.basescan.org/tx/0x5709f71f9b27c6ed083d816265085ef812eeee4d114db7c96956287bb0df0539

Run the Demo Script

make demo-sepolia

Testing and Coverage

100% Pass Rate Proof

• forge test: 38/38 passing (root contracts)
• forge test --root reactive: 8/8 passing (reactive package)
• Combined Forge pass rate: 46/46 passing

Coverage Proof (latest run)

• Command used: cd contracts && forge coverage --ir-minimum --report summary
• Total coverage from latest report:
  • Lines: 73.83% (347/470)
  • Statements: 71.00% (377/531)
  • Branches: 42.72% (44/103)
  • Functions: 92.31% (60/65)
• Critical contract highlights:
  • UniBrainTelemetryHook.sol: 100% lines, statements, branches, funcs
  • UniBrainStrategyBook.sol: 97.50% lines, 100% branches, 100% funcs
  • UniBrainStrategyExecutor.sol: 100% branches, 100% funcs
  • UniBrainReactiveStrategy.sol: 98.90% lines/statements, 100% funcs

Test Categories Implemented

• Unit tests
• Edge-case tests
• Fuzz tests
• Integration lifecycle tests
• Callback authentication / replay tests
• Subscription reliability tests (constructor-no-subscribe, strict activate, failure reverts, health views)

Future Roadmap

• Raise total branch/line coverage to full 100% by expanding branch-focused tests in shared AI and deploy script paths.
• Add destination execution assertion path to demo for deterministic DecisionExecutionAttempt/state-diff proof in one run.
• Add richer frontend observability for per-decision provenance, nonce tracking, and replay diagnostics.
• Add gas benchmark gates in CI for hook and executor critical paths.
• Expand multi-pool/multi-regime stress testing and invariant suites.