Uniswap logo

Skill

permissioned-pools-configurator

configure Uniswap v4 permissioned pools

Published by Uniswap Updated Aug 14
Covers Configuration DeFi Smart Contracts Ethereum

Description

Interactively collect and validate the parameters for a Uniswap v4 Permissioned Pool setup - chain, underlying token, allowlist checker, adapter owner, verification deposit, the four wrapper registrations, hook, and pool pricing - and emit the JSON config the deployer skill consumes. Use when the user says "configure a permissioned pool", "permissioned pool config", "adapter parameters", "permissioned pool setup parameters", "PermissionsAdapter configuration", or asks to set up parameters for a permissioned pool before deploying. Covers contract mechanics only - does not constitute legal, financial, investment, tax, or compliance advice, and is not a compliance review.

SKILL.md

Permissioned Pools Configurator

Interactive bulk-form configurator for a Uniswap v4 Permissioned Pool. Collects every parameter the setup journey needs, validates each one, and displays a single JSON configuration object that the permissioned-pools-deployer skill consumes to run the on-chain sequence. For the contract mechanics behind any of these parameters, see permissioned-pools-issuer.

Runtime Compatibility: This skill uses AskUserQuestion to collect parameters in batches. If AskUserQuestion is not available in your runtime, collect the same parameters through natural language conversation instead, in the same batch groupings and with the same validation after each group.

Scope and Disclaimer

This is a configuration tool, not deployment guidance and not a compliance review.

  • ✅ It collects and validates the inputs a permissioned-pool setup needs and produces a JSON document. It does not call any contract, does not broadcast anything, and holds no signing keys.
  • Treat this as contract mechanics only. This skill does not constitute legal, financial, investment, or tax advice, and it is not a compliance review of your token, your allowlist, your KYC or AML program, or your configuration. It covers what each parameter means and what breaks if it is wrong.
  • ✅ It never invents a contract address. Addresses you supply are recorded as given; addresses you do not yet have — whether because they still need to be looked up against the deploy guide's table or because the issuer has not decided on them yet — are stored as the literal string "RESOLVE", never guessed or recalled from memory.
  • ✅ Review the emitted configuration yourself, and re-resolve every address it marks for resolution, before handing it to a deployment flow.
  • Read the repo's usage guidelines. The repo root DISCLAIMER.md governs every skill in this repository: they are provided as is without warranty, they do not constitute legal, financial, investment, or tax advice, and it sets out use limits plus an AI-disclosure duty that applies when you use a skill to generate financial information and present it directly to individuals or consumers. Point the user to it.

State this framing inline (quoted or in your own words) and continue with the configuration flow in the same response — this is reference and data-collection content, not an action-oriented deployment step, so no acknowledgment gate is needed here. The deployer skill gates on acknowledgment before anything gas-spending.

What This Config Is (and Isn't)

The output is a pre-deployment plan, not a record of what already exists on-chain.

  • The PermissionsAdapter address does not appear in this config. It does not exist until the deployer's Step 2 (createPermissionsAdapter) creates it. If the issuer already has a deployed adapter for this token, that is a re-run/import scenario this skill does not currently handle — collect its address as a note and confirm with the deployer skill's preflight checks instead of adding it here. This declines only the adapter address itself — it does not block the rest of the request. Continue validating and including every other field the user supplied (or marked "RESOLVE") in the emitted configuration exactly as the flow below describes.
  • The underlying permissioned token is the only token address this config records. See the two-addresses distinction in permissioned-pools-issuer — the pool currency the deployer builds later is the adapter, derived at deploy time, not configured here.
  • Wrapper and hook addresses are chain-specific and change over time. This skill asks whether you already have them resolved for your target chain; if not, it marks them for resolution rather than filling in a value.

Configuration Flow

Collect parameters in four batches of at most four questions each. After each batch, validate every answer against config-schema.md before moving on, and show a running summary of what has been collected and what remains. Field-by-field meaning, how to choose each value, and what breaks if it is wrong are all in parameter-reference.md.

Front-loaded or skip-ahead answers

A user may answer several batches' worth of questions in one free-form message, or ask to skip the remaining questions and finish immediately. Do not withhold the whole configuration while asking whether to proceed. Instead: validate everything the user did supply against config-schema.md, apply the literal string "RESOLVE" (or null for allowlistChecker.address when the checker is to-be-deployed) to every address-type field left open, and produce that partial configuration in the same response. Fields with no "RESOLVE" sentinel — the verification deposit amount, pool currency, fee tier, tick spacing, starting price, and seeding — have no safe default; flag each one explicitly as still needing an answer rather than guessing at it, and ask only about those. This includes the verification deposit amount even though Batch 2 lists "1 wei" as the recommended option: a menu option a user has not actually picked is not a value the skill may silently apply on their behalf — treat an unanswered verification deposit amount exactly like the other sentinel-less fields, never as an already-resolved "1". Skipping the batch-by-batch confirmation step never means skipping validation — every rule above still applies to whatever the user did supply.

Batch 1: Network, Token & Ownership (4 questions)

Question 1 — Network

  • Prompt: "Which network is this permissioned pool for?"
  • Options: "Ethereum Mainnet (chain ID 1)", "Unichain (chain ID 130)", "Sepolia (chain ID 11155111)", Custom chain ID (via "Other")
  • Note in the prompt: availability of the permissioned-pools contracts on a given chain is not something this skill asserts — confirm the chain has a row in the deploy guide's #deployment-addresses table before proceeding.
  • Store: chainId

Question 2 — Underlying permissioned token

  • Prompt: "What is the address of the permissioned ERC-20 being listed?"
  • Options: "Not deployed yet" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Validation: ^0x[a-fA-F0-9]{40}$, or the literal string "RESOLVE" if not yet deployed
  • Store: permissionedToken

Question 3 — Allowlist checker

  • Prompt: "Do you already have an allowlist checker deployed for this token?"
  • Options, each mapped to the exact allowlistChecker value it produces:
    • "Yes, use an existing address" (then ask for the address via "Other") → mode: "existing", address set to that address.
    • "No, I still need to deploy one" → mode: "to-be-deployed", address set to null.
  • Validation: address form above when "existing" is chosen
  • Store: allowlistChecker.mode (existing | to-be-deployed), allowlistChecker.address

Question 4 — Adapter owner

  • Prompt: "What address should own the PermissionsAdapter (initialOwner)?"
  • Options: "Not yet decided — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Note in the prompt: this key can change the allowlist checker, add or remove wrappers, pause and resume swapping, and force-exit any LP position — the published guide recommends a multisig. Do not pick this in a hurry just to finish the flow; use the resolution marker instead if the multisig or key is not decided yet.
  • Validation: ^0x[a-fA-F0-9]{40}$ and not the zero address, or the literal string "RESOLVE"
  • Store: adapterOwner

Validate: confirm chain ID is a positive integer, the token address is either a valid address or the literal string "RESOLVE", the checker address (if "existing" was chosen) matches the address regex, and the adapter owner is either a valid non-zero address or the literal string "RESOLVE". Show a summary of the four collected values, flagging any field still marked "RESOLVE".


Batch 2: Verification & Core Wrappers (4 questions)

Question 1 — Verification deposit amount

  • Prompt: "How much of the underlying token should be deposited for verification?"
  • Options: "1 wei (recommended)", Custom amount (via "Other")
  • Note in the prompt: this deposit becomes the adapter's mintable headroom, there is no withdraw function, and 1 wei is sufficient — see Parameter Reference before choosing anything larger.
  • "Recommended" describes the value if asked and picked — it is never a license to fill in "1" on the user's behalf when this question was skipped. See Front-loaded or skip-ahead answers above.
  • Store: verificationDepositAmount

Question 2 — PermissionedPositionManager address

  • Prompt: "Do you have the PermissionedPositionManager address resolved for this chain?"
  • Options: "Not yet resolved — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Validation: address form, or the literal string "RESOLVE"
  • Store: allowedWrappers.permissionedPositionManager

Question 3 — Universal Router (v2.2+) address

  • Prompt: "Do you have the Universal Router #v2.2 (or higher) address resolved for this chain?"
  • Options: "Not yet resolved — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Note in the prompt: the plain UniversalRouter deployments key is a different, non-permissioned router — only the #v2.2 deployment works here.
  • Validation: address form, or the literal string "RESOLVE"
  • Store: allowedWrappers.universalRouterV2_2

Question 4 — PermissionedHooks address

  • Prompt: "Do you have the PermissionedHooks address resolved for this chain, for setAllowedHook?"
  • Options: "Not yet resolved — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Note in the prompt: this is the shared hook address, registered with setAllowedHook on the position manager — it is never registered as a wrapper.
  • Validation: address form, or the literal string "RESOLVE"
  • Store: hook

Validate: confirm the deposit amount is a positive integer (not zero) expressed in the token's smallest unit, and that every address field is either a valid address or the literal string "RESOLVE" — never a guessed or invented value. Show a summary.


Batch 3: Remaining Wrappers & Pool Currency (4 questions)

Question 1 — V4Quoter address

  • Prompt: "Do you have the V4Quoter address resolved for this chain?"
  • Options: "Not yet resolved — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Validation: address form, or the literal string "RESOLVE"
  • Store: allowedWrappers.v4Quoter

Question 2 — MixedRouteQuoterV2 address

  • Prompt: "Do you have the MixedRouteQuoterV2 address resolved for this chain?"
  • Options: "Not yet resolved — mark for resolution" (stores the literal string "RESOLVE"), custom address (via "Other")
  • Validation: address form, or the literal string "RESOLVE"
  • Store: allowedWrappers.mixedRouteQuoterV2

Question 3 — Paired currency

  • Prompt: "What currency should the pool pair against the permissioned token?"
  • Options: "Native ETH", custom ERC-20 address (via "Other")
  • Note in the prompt: the permissioned token's side of the pool is always the adapter, never the underlying token — this question is only about the other currency.
  • Validation: address form for an ERC-20 and not the zero address, or the literal string "native" (the native-currency sentinel; see Address Validation for the full semantics). A user who supplies the zero address here means "native" — record "native", never the zero address.
  • Store: pool.pairedCurrency

Question 4 — Fee tier

  • Prompt: "What fee tier should the pool use?"
  • Options: "0.05% (500)", "0.30% (3000)", "1.00% (10000)", Custom (via "Other")
  • Validation: positive integer, hundredths of a basis point
  • Store: pool.feeTier

Validate: confirm both remaining wrapper fields are a valid address or the literal string "RESOLVE", the paired currency is a valid non-zero address or the literal string "native", and the fee tier is a positive integer. When the paired currency is a resolved address and permissionedToken is also a resolved address, confirm the two differ — pairing the permissioned token against itself builds a PoolKey that either fails at the PoolManager level or creates a pool nobody intended (see Parameter Reference). Show a summary of all four wrappers, the hook, and the pool currency collected so far, and flag every field still marked "RESOLVE".


Batch 4: Pricing & Seeding (3 questions)

Question 1 — Tick spacing

  • Prompt: "What tick spacing should the pool use?"
  • Options: "10 (stable-like pairs)", "60 (standard)", "200 (wide-range pairs)", Custom (via "Other")
  • Validation: positive integer
  • Store: pool.tickSpacing

Question 2 — Starting price

  • Prompt: "What should the pool's starting price be, expressed as paired-currency per permissioned-token?"
  • Options: "1:1 (equal value ratio)", Custom ratio (via "Other")
  • Validation: positive number
  • Store: pool.startingPriceRatio (a human-readable ratio; the deployer skill converts this to sqrtPriceX96 at initialization time, since that conversion depends on the decimals of both currencies, which are not yet resolved at configuration time)

Question 3 — Seeding intent

  • Prompt: "Do you plan to seed liquidity as part of this same setup?"
  • Options, each mapped to the exact seeding.intent value it produces:
    • "Yes, from the adapter owner's wallet" → intent: "seed-now", wallet set to the same address already collected as adapterOwner in Batch 1. Offer this option only when adapterOwner is a real address — not when it is the literal string "RESOLVE". seeding.wallet must always be a real address (see Validation below), so this option cannot copy forward an unresolved owner.
    • "Yes, from a different wallet" (then ask for the address via "Other") → intent: "seed-now", wallet set to that address
    • "Not yet — pool creation only" → intent: "none", wallet: null
  • If adapterOwner is "RESOLVE", present only the second and third options — drop the first option from the list entirely rather than showing it disabled.
  • Note in the prompt: whichever wallet seeds liquidity, and the wallet that will receive the position, both need LIQUIDITY_ALLOWED on the issuer's own allowlist before the first mint — being the adapter owner grants nothing on the liquidity path.
  • Validation: when intent is "seed-now", wallet is required and must match the address regex — never the literal string "RESOLVE" or any other sentinel; when intent is "none", wallet must be null. This flow never produces "seed-later" — that enum value exists in the schema for a config edited or extended outside this flow, not for output from these three options.
  • Store: seeding.intent, seeding.wallet

Validate: confirm tick spacing is a positive integer, the starting price ratio is a positive number, and — per the branch taken above — either seeding.wallet is a valid, non-"RESOLVE" address (when intent is "seed-now") or it is null (when intent is "none"). seeding.wallet never accepts the literal string "RESOLVE", even when it was going to be copied from adapterOwner — see the option restriction above. Show the full configuration collected across all four batches.


Generate and Display the Configuration

Assemble everything collected into the JSON shape defined in config-schema.md, keyed by chainId at the top level:

{
  "<chainId>": {
    "permissionedToken": "...",
    "allowlistChecker": { "mode": "...", "address": "..." },
    "adapterOwner": "...",
    "verificationDepositAmount": "...",
    "allowedWrappers": {
      "permissionedPositionManager": "...",
      "universalRouterV2_2": "...",
      "v4Quoter": "...",
      "mixedRouteQuoterV2": "..."
    },
    "hook": "...",
    "pool": {
      "pairedCurrency": "native",
      "feeTier": 3000,
      "tickSpacing": 60,
      "startingPriceRatio": 1
    },
    "seeding": { "intent": "seed-now", "wallet": "..." }
  }
}

Display this JSON directly in the transcript. Do not automatically create a file. Let the user copy it or choose a filepath in the next-steps question below.

Display Summary

Immediately after the JSON, show a human-readable summary:

  • Network and chain ID.
  • Underlying token and allowlist-checker mode.
  • Adapter owner.
  • Verification deposit amount, with a one-line reminder that it is not recoverable.
  • Every wrapper and the hook, each flagged resolved or needs resolution.
  • Pool currency, fee tier, tick spacing, starting price ratio.
  • Seeding intent and wallet, with a reminder that the wallet needs LIQUIDITY_ALLOWED.
  • A validation checklist: which rules from config-schema.md passed, and which fields still need a real address before this config can be used to deploy anything.

Next Steps

Ask the user what they want to do:

  • "Save to file" — ask for a filepath, default permissioned-pool-config.json. Require a relative path ending in .json; reject an absolute path (leading /), reject a leading ~ (a runtime that expands it can write outside the working directory — including over Claude Code's own config store), and reject any path containing a .. segment. Resolve the path and confirm it still stays within the current working directory before writing. If the resolved path already exists, ask for confirmation before overwriting it — including a file that looks unrelated to this skill, such as package.json or tsconfig.json. Re-prompt once on a rejected path rather than silently substituting the default. Then write exactly the JSON object shown above. (This validation is intentionally stricter than other skills' "save to file" steps in this repo — the ~ and .. rejections were added in response to a security audit finding that a leading ~ could target ~/.claude.json, so keep them here rather than trimming to match a lighter-weight equivalent elsewhere.)
  • "View the setup walkthrough" — point at permissioned-pools-deployer.
  • "Modify configuration" — re-run the relevant batch above.
  • "Exit" — end here; the user can copy the JSON from the transcript.

Notes for Implementers of This Flow

  • Never fill in a token, wrapper, hook, or adapter-owner address from memory. If the user does not supply one, store the exact literal string "RESOLVE" in that field — full stop. This includes not "helpfully" completing a partially-typed address.
  • Batch, validate, summarize — every time. Do not run all four batches back to back without the validation and summary steps in between; that is what keeps a bad input from propagating into the next batch's defaults.
  • This skill does not read chain state, fetch an RPC, or consult a deployments JSON file. It has no Bash or WebFetch in its tool list, on purpose — resolving addresses against Uniswap/contracts deployments/json/<chainId>.json and the deploy guide's table is the user's job, or a later step in the deployer skill's preflight.

© 2026 YourAI.tools. Every skill from an identity-verified publisher.

Independent catalog. Not affiliated with, endorsed by, or sponsored by Anthropic or any listed publisher. All trademarks belong to their respective owners.