Skip to content
cccskills
FreePublished on GitHub

svm-warp-alt-manage

Create, check, and read on-chain Sealevel (SVM) Address Lookup Tables (ALTs) for a warp route. Use after deploying, extending, or upgrading an SVM warp route, when transfers fail with transaction-too-large errors, or when asked to inspect or verify a route's ALTs. Fee-enabled SVM transfers bundle 40+ accounts and require ALTs to fit under the 1232-byte tx limit.

The English summary is being prepared. Showing the original description.

Run score Not tested

See how to install

Files included (1)

  • SKILL.md5.8 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

SVM Warp ALT Management

Manage the on-chain Address Lookup Tables (ALTs) that a Sealevel warp route uses to compress its account list into a v0 VersionedTransaction. Fee-enabled SVM warp transfers reference 40+ accounts (core + fee + IGP + token plugin) and exceed Solana's 1232-byte transaction size limit without ALTs.

Wraps the hyperlane warp alt command group (typescript/cli/src/commands/warp-alt.ts): create, check, read.

When to Use

  • Right after deploying a new SVM warp route (/warp-deploy-init-route), extending one, or upgrading an SVM program (/svm-warp-program-upgrade) — ALTs must be (re)created.
  • SVM transfers fail with transaction-too-large / too-many-accounts errors.
  • "check the ALTs for <route>" / "are the lookup tables up to date?" → check.
  • "show me the ALT contents for <route>" → read.

Model

Each Sealevel chain in the route has, in the registry warp config under options.sealevel.altAddresses.<chain>:

  • core — one chain-shared ALT (mailbox/core accounts, reusable across routes on that chain).
  • warpSpecific — an array (min 1) of route-specific ALTs (this route's token/fee/IGP accounts).

ALTs are created on-chain, populated, then frozen. Frozen tables are immutable and their rent cannot be reclaimed — you never "edit" an ALT, you create a fresh one and update the registry pointer. This is why --force/--full-force leak the old (frozen) tables: they create new ones and abandon the old, unrecoverable ones.

Commands

All commands run from typescript/cli and require --warp-route-id. --chain is optional and scopes to a single chain (defaults to all Sealevel chains in the route). create needs write context (--key.sealevel + private RPC via HTTP registry); check and read are read-only.

create — build + persist ALTs

cd <MONOREPO_ROOT>/typescript/cli && pnpm --silent hyperlane warp alt create \
  --registry http://localhost:<port> \
  --key.sealevel "$SEALEVEL_KEY_VAR" \
  -w <WARP_ROUTE_ID> \
  [--chain <chain>]
  • Default (no flags): creates ALTs only where the registry has none. Existing entries are left as-is.
  • --force / -f: recreate the warp-specific ALTs for chains that already have registry entries; the chain-shared core ALT is reused. Old warp-specific frozen ALTs are leaked (unrecoverable).
  • --full-force / -F: recreate all ALTs including core. Implies --force. All old frozen ALTs leaked.
  • Writes the resulting ALT addresses back to the registry warp config (options.sealevel.altAddresses).

Use plain create after a fresh deploy. Use --force after an upgrade/extend that changed the route-specific account set. Reserve --full-force for when the core ALT itself is wrong — it's the most wasteful (leaks the shared table too).

check — verify on-chain ALTs match expected

cd <MONOREPO_ROOT>/typescript/cli && pnpm hyperlane warp alt check \
  --registry http://localhost:<port> \
  -w <WARP_ROUTE_ID> \
  [--chain <chain>]

Compares on-chain ALT contents against the expected account set for the route. Exits non-zero on drift — treat a non-zero exit as "ALTs are stale, recreate them". check reports core and warpSpecific drift separately, and the fix differs by bucket: recreate warpSpecific drift with create --force (reuses the shared core ALT); recreate core drift (including an unfrozen core) with create --full-force. Do NOT use --force for a core diff — it reuses the existing core, so the core stays broken and the next check still fails. Report the specific diff and its bucket to the user.

read — dump ALT contents

cd <MONOREPO_ROOT>/typescript/cli && pnpm hyperlane warp alt read \
  --registry http://localhost:<port> \
  -w <WARP_ROUTE_ID> \
  [--chain <chain>] \
  [--out <file.yaml>]

Read-only. Prints (or writes) the current on-chain ALT addresses and their entries. No key needed.

Execution Flow

  1. Start the HTTP registry per /start-http-registry (needed for private Sealevel RPC; add --writeMode for create, omit for check/read). Note the port + task ID.
  2. Run the requested subcommand (create / check / read) against --registry http://localhost:<port>.
  3. For create: after success, confirm the registry warp config now has options.sealevel.altAddresses.<chain>.{core,warpSpecific} and follow with a check to prove no drift.
  4. Stop the HTTP registry per /stop-http-registry, even on failure.
  5. If create mutated the registry, open a registry PR with the updated warp config (the altAddresses block) so the ALT pointers are canonical.

Caveats

  • ALTs are frozen and unrecoverable. Never assume you can reclaim rent or mutate a table. --force/--full-force intentionally abandon old frozen tables — only use them when the account set actually changed.
  • Always create ALTs after any deploy/extend/upgrade of an SVM route. Missing or stale ALTs are the usual cause of transaction-too-large failures on fee-enabled SVM transfers.
  • check exit code is the signal — non-zero = drift; don't ignore it. Recreate and re-check.
  • Core vs warp-specific: match the fix to the drift bucket check reports — --force for warpSpecific drift (keeps the shared core ALT), --full-force for core drift. A --force won't touch the core, so using it on a core diff leaves the check failing.
  • create needs write context and a funded --key.sealevel; check/read do not.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Write a changeset file in the `hyperlane-registry` repo for a PR that adds, updates, or modifies warp route config / chain metadata / addresses. Canonical pattern — never run the interactive `pnpm changeset` CLI; write the file directly. Referenced by any skill that opens a registry PR.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

Trigger Docker image builds for Hyperlane agent, monorepo, or node service images. Use when the user wants to build new Docker images for a branch, commit, or tag.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

The low-level on-chain probes for identifying what an owner address is — cast code → EOA / EIP-7702-delegated EOA / contract, VERSION() → Gnosis Safe, read-only get-owner-ica.ts derive → ICA derivation, getAccountInfo → Sealevel account existence — with how to interpret each result. The primitive layer beneath validate-owners / resolve-artifacts / update-owners, which each apply their own per-chain or per-artifact policy on top.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

Review code changes using Hyperlane monorepo coding standards. Use when reviewing PRs, checking your own changes, or doing self-review before committing.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

Debug why a Hyperlane message is not being processed. Use when given a message ID or explorer URL to investigate delivery failures, gas estimation errors, validator issues, or other processing problems.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

Debug RPC provider health issues for a chain. Use when alerts mention RPC failures, high latency, stale blocks, or provider errors. Diagnoses the issue and recommends fixes.

The English summary is being prepared. Showing the original description.

hyperlane-xyz/hyperlane-monorepo76Updated Oct 23, 2026

See all skills by hyperlane-xyz

Report a problem with this skill