V4 Hook Generator logo

V4 Hook Generator

Organization
Uniswap
v4-hook-generator

Generate Uniswap v4 hook contracts via OpenZeppelin MCP. Use when building custom swap logic, async swaps, hook-owned liquidity, custom curves, dynamic fees, MEV protection, limit orders, or oracle hooks.

Overview

PublisherUniswap
Repositoryuniswap-ai
Skill namev4-hook-generator
Stars
234
Forks
40
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Uniswap on GitHub. Read the source before you install it.

Installation

Install the V4 Hook Generator AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Uniswap/uniswap-ai/tree/main/packages/plugins/uniswap-hooks/skills/v4-hook-generator
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Uniswap/uniswap-ai.git /tmp/uniswap-ai
mkdir -p .claude/skills
cp -r /tmp/uniswap-ai/packages/plugins/uniswap-hooks/skills/v4-hook-generator .claude/skills/v4-hook-generator
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable V4 Hook Generator in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads V4 Hook Generator on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so V4 Hook Generator is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

v4 Hook Generator

Generate Uniswap v4 hook contracts via the OpenZeppelin Contracts Wizard MCP tool. This skill guides you through selecting the right hook type, configuring permissions and utilities, assembling the canonical MCP JSON, and invoking the MCP tool to produce ready-to-compile Solidity code.

Security companion: Generated hook code touches fund-handling contracts. Always apply the v4-security-foundations skill immediately after generation to audit permissions, delta accounting, and access control before deploying to any network.

When to Use This Skill

Use this skill when you need to:

  • Scaffold a new Uniswap v4 hook contract from scratch
  • Select the right base hook type for a specific use case (fees, MEV protection, oracles, etc.)
  • Configure hook permissions, utility libraries, shares, and access control
  • Produce the canonical MCP tool call JSON to invoke the OpenZeppelin Contracts Wizard
  • Understand trade-offs between hook configuration options before committing to an implementation

Prerequisite / companion skill: v4-security-foundations — run it before writing custom logic and again before deployment. Hook misconfiguration can drain user funds.

Hook Type Decision Table

Choose the base hook type that matches your primary goal. If your hook has multiple goals, choose the type that covers the most critical concern and layer additional logic on top.

GoalUse Hook
Custom swap logicBaseHook
Async/delayed swapsBaseAsyncSwap
Hook-owned liquidityBaseCustomAccounting
Custom curveBaseCustomCurve
Dynamic LP feesBaseDynamicFee
Dynamic swap feesBaseOverrideFee
Post-swap feesBaseDynamicAfterFee
Fixed hook feesBaseHookFee
MEV protectionAntiSandwichHook
JIT protectionLiquidityPenaltyHook
Limit ordersLimitOrderHook
Yield on idleReHypothecationHook
OracleBaseOracleHook
V3-compatible oracleOracleHookWithV3Adapters

Selection tips:

  • BaseHook is the general-purpose starting point — choose a specialized type only when the built-in logic provides concrete value.
  • BaseCustomCurve replaces the entire AMM math; only use it if you are implementing a novel pricing algorithm.
  • AntiSandwichHook and LiquidityPenaltyHook both address MEV but target different actors (traders vs. JIT LPs). Clarify which attack vector you are mitigating.
  • OracleHookWithV3Adapters is appropriate when downstream integrations expect a Uniswap v3 IUniswapV3Pool-compatible oracle interface.

Minimal Decision Checklist

Before calling the MCP tool, confirm all six decisions:

  1. Hook type — chosen from the decision table above
  2. Permissions to enable — only the callbacks your logic actually uses (beforeSwap, afterSwap, etc.)
  3. Utility libraries — currencySettler, safeCast, transientStorage as needed
  4. Shares — false, ERC20, ERC6909, or ERC1155
  5. Access control — ownable, roles, or managed
  6. Hook inputs — blockNumberOffset, maxAbsTickDelta (only for hook types that use them)

Permission Configuration

All 14 permission flags with guidance on when to enable each. Start with all flags false and enable only what your hook logic requires. Every enabled permission increases the hook's attack surface and requires a specific bit to be set in the hook's deployed address (see address encoding note below).

Permission FlagEnable WhenRisk
beforeInitializeYou need to validate or restrict pool creation paramsLOW
afterInitializeYou need to set up state after a pool is createdLOW
beforeAddLiquidityYou need to gate or transform LP depositsMEDIUM
afterAddLiquidityYou track LP positions or distribute rewardsLOW
beforeRemoveLiquidityYou need lock-up periods or fee-on-exit logicHIGH
afterRemoveLiquidityYou track position removals for accountingLOW
beforeSwapYou modify swap behavior, apply dynamic fees, or blockHIGH
afterSwapYou observe final swap state for oracles or accountingMEDIUM
beforeDonateYou restrict who may donate to the poolLOW
afterDonateYou track donation eventsLOW
beforeSwapReturnDeltaYou implement custom AMM curves or JIT liquidity (CRITICAL: NoOp attack vector — see v4-security-foundations)CRITICAL
afterSwapReturnDeltaYou extract a hook fee from swap outputHIGH
afterAddLiquidityReturnDeltaYou adjust LP token amounts on depositHIGH
afterRemoveLiquidityReturnDeltaYou adjust withdrawal amountsHIGH

Address encoding note: Permissions are encoded as bits in the hook contract's deployed address. The address must have the correct bits set at deployment time or the PoolManager will revert. Use HookMiner (from v4-periphery) to mine a salt that produces an address with the correct bit pattern. Never change permissions after deployment — the address is immutable.

Utility Library Selection

Three optional utility libraries can be included in the generated hook. Include only what your hook logic uses.

LibraryInclude When
currencySettlerYour hook moves tokens between itself and the PoolManager (e.g., custom accounting, fee collection)
safeCastYour hook performs arithmetic that could overflow when casting between integer types
transientStorageYour hook needs to pass data between callbacks within a single transaction without persisting to storage (requires EVM Cancun or later, Solidity >= 0.8.24)

Guidance:

  • currencySettler is almost always needed when beforeSwapReturnDelta, afterSwapReturnDelta, afterAddLiquidityReturnDelta, or afterRemoveLiquidityReturnDelta are enabled — it provides settle and take helpers that implement the correct sync → transfer → settle sequence.
  • transientStorage is a gas-efficient alternative to storage slots for intra-transaction state. Use it to pass a flag or value from beforeSwap to afterSwap without paying 20k gas for a cold SSTORE.
  • safeCast is advisable whenever you compute amounts derived from int256/uint256 conversions, especially for fee calculations.

Shares Configuration

The shares option controls whether the generated hook issues a token representing user shares (e.g., LP positions in hook-owned liquidity pools).

OptionDescriptionUse When
falseNo share token — hook does not track ownership of deposited assetsSimple hooks that do not hold user funds
ERC20Fungible share token — one token represents proportional ownership of all hook-held assetsHook-managed liquidity pools with interchangeable shares
ERC6909Multi-token (minimal) — one contract manages many token IDs with lower overhead than ERC1155Hook manages multiple distinct asset classes efficiently
ERC1155Multi-token (standard) — full ERC1155 with metadata URI supportHook needs broad wallet and marketplace compatibility

Trade-offs:

  • false: smallest bytecode, no share accounting overhead; appropriate for fee hooks and oracles.
  • ERC20: simplest fungible share; good DeFi composability (e.g., used as collateral).
  • ERC6909: gas-efficient multi-token with a minimal interface; preferred for new protocol designs.
  • ERC1155: widest ecosystem support (wallets, explorers, NFT marketplaces); higher gas cost per transfer than ERC6909.

Access Control Options

The access option shapes the constructor and administrative interface of the generated hook.

OptionConstructor ShapeUse When
ownableconstructor(IPoolManager, address initialOwner)Single owner controls all admin functions
rolesconstructor(IPoolManager, address admin)Multiple roles with granular permissions (OpenZeppelin AccessControl)
managedconstructor(IPoolManager, address authority)External authority contract governs permissions (OpenZeppelin AccessManaged)

Guidance:

  • ownable is the simplest — one address can perform all privileged operations. Suitable for early-stage hooks and personal tools.
  • roles adds ADMIN_ROLE, PAUSER_ROLE, etc. via OpenZeppelin AccessControl. Use when different team members need different privileges (e.g., a keeper bot that can update fees but cannot upgrade the contract).
  • managed delegates all permission checks to a separate AccessManager contract. Use when you need a unified governance layer across multiple contracts or want timelocked admin actions.

Note: Changing the access option changes the constructor signature. Update deployment scripts and initialization logic accordingly. When using ownable, ensure the initialOwner is not the zero address — OpenZeppelin's Ownable reverts on zero address since v5.

Hook Inputs Reference

Some hook types accept numeric configuration inputs that tune behavior. These are passed as the inputs object in the MCP tool call.

InputTypeUsed ByDescription
blockNumberOffsetuint256AntiSandwichHook, LiquidityPenaltyHookNumber of blocks before sandwich/JIT detection window opens
maxAbsTickDeltaint24AntiSandwichHookMaximum tick movement allowed per block before MEV protection triggers

For hook types that do not use these inputs, omit the inputs field or pass an empty object {}. Passing unsupported inputs to the MCP tool will not cause an error but the values will be ignored.

MCP Tool Call (Canonical)

The OpenZeppelin Contracts Wizard exposes a generate_hook MCP tool. The following is the canonical JSON schema — populate each field according to your decisions from the sections above, then pass this object as the tool's argument.

json
{
  "hook": "BaseHook",
  "name": "MyHook",
  "pausable": false,
  "currencySettler": true,
  "safeCast": true,
  "transientStorage": false,
  "shares": { "options": false },
  "permissions": {
    "beforeInitialize": false,
    "afterInitialize": false,
    "beforeAddLiquidity": false,
    "beforeRemoveLiquidity": false,
    "afterAddLiquidity": false,
    "afterRemoveLiquidity": false,
    "beforeSwap": true,
    "afterSwap": false,
    "beforeDonate": false,
    "afterDonate": false,
    "beforeSwapReturnDelta": false,
    "afterSwapReturnDelta": false,
    "afterAddLiquidityReturnDelta": false,
    "afterRemoveLiquidityReturnDelta": false
  },
  "inputs": {
    "blockNumberOffset": 1,
    "maxAbsTickDelta": 100
  },
  "access": "ownable",
  "info": { "license": "MIT" }
}

Field notes:

  • hook: string — one of the 14 hook types from the decision table
  • name: string — the Solidity contract name (PascalCase, no spaces)
  • pausable: boolean — wraps the hook in OpenZeppelin Pausable; adds pause()/unpause() admin functions
  • shares.options: false | "ERC20" | "ERC6909" | "ERC1155"
  • access: "ownable" | "roles" | "managed"
  • info.license: SPDX license identifier — use "MIT" for open-source hooks
  • inputs: omit or pass {} for hook types that do not use blockNumberOffset/maxAbsTickDelta

Workflow: Gather → Configure → Generate → Secure

Follow these steps in order every time you use this skill.

Step 1: Gather Requirements

Ask the user (or infer from context):

  1. What is the hook's primary goal? (Map to the decision table.)
  2. Which lifecycle events does the hook need to intercept? (Map to permissions.)
  3. Does the hook hold or move user funds? (Determines currencySettler and shares.)
  4. Who administers the hook? (Single owner, role-based team, or external governance?)
  5. Does the hook need to pass state between callbacks within a single transaction?
  6. Is this for a chain with EVM Cancun support? (Required for transientStorage.)

Step 2: Select Hook Type

Using the decision table, identify the single best hook type. If the user's goal maps to multiple types, explain the trade-offs and ask them to confirm. Document the chosen type and the reasoning.

Step 3: Configure All Six Decisions

Work through the minimal decision checklist:

  1. Set hook to the chosen type.
  2. Set each permission flag — default false, enable only what the logic requires.
  3. Set currencySettler, safeCast, transientStorage based on utility library guidance.
  4. Set shares.options based on shares guidance.
  5. Set access based on access control guidance.
  6. Set inputs only if the hook type uses blockNumberOffset or maxAbsTickDelta.

Step 4: Assemble and Call the MCP Tool

Construct the JSON object from Step 3 and call the OpenZeppelin Contracts Wizard MCP tool with it. The tool returns Solidity source code — it does not write files automatically.

After receiving the generated code:

  1. Display the code to the user.
  2. Explain the key generated sections (constructor, getHookPermissions, enabled callbacks).
  3. Note any manual steps required (HookMiner for address mining, deployment script updates for constructor args if access is roles or managed).

Step 5: Apply Security Foundations

Always remind the user — and invoke v4-security-foundations — before the code is deployed:

  • Verify all enabled callbacks check msg.sender == address(poolManager).
  • Review any enabled *ReturnDelta permissions for NoOp attack exposure.
  • Confirm delta accounting sums to zero for every execution path.
  • Run the full pre-deployment audit checklist from v4-security-foundations.

Important Notes

  • Access control changes constructor shape: Choosing ownable adds an initialOwner parameter; roles adds an admin parameter; managed adds an authority parameter. Update deployment scripts and factory contracts accordingly.
  • Permissions encode in the hook address: Each enabled permission flag corresponds to a specific bit in the lower bytes of the hook's deployed address. The PoolManager validates these bits on every callback. Use HookMiner from v4-periphery to mine a deployment salt that produces a matching address.
  • MCP returns code only — it does not write files: The generated Solidity is returned as a string. You must write it to disk yourself (e.g., packages/contracts/src/hooks/MyHook.sol).

Related Skills

  • v4-security-foundations — Run this after generation. Security audit for Uniswap v4 hooks: permission risk matrix, NoOp attack patterns, delta accounting, access control verification, and the full pre-deployment audit checklist. Generated hook code should never be deployed without completing this review.
  • viem-integration — Deploy generated hook contracts and interact with them using viem/wagmi
  • v4-sdk-integration — Interact with deployed hooks via the Uniswap v4 SDK

Frequently asked questions

What does the V4 Hook Generator AI skill do?

Generate Uniswap v4 hook contracts via OpenZeppelin MCP. Use when building custom swap logic, async swaps, hook-owned liquidity, custom curves, dynamic fees, MEV protection, limit orders, or oracle hooks.

Why use V4 Hook Generator on TypingMind?

Because you install it once and use it with any model. V4 Hook Generator is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install V4 Hook Generator in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Uniswap/uniswap-ai/tree/main/packages/plugins/uniswap-hooks/skills/v4-hook-generator. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use V4 Hook Generator?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with V4 Hook Generator?

As many as you like. As long as a model supports skills, you can use V4 Hook Generator with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the V4 Hook Generator AI skill free?

Yes. It is published on GitHub by Uniswap under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇