We manage your security to weather the AI storm →
Get a quote

Liquid staking audit

A liquid staking token promises its holder the staked asset back, at a rate the contracts compute, through a queue that waits on validators. Guardian audits the contracts behind that promise, with public staking reports on Ethereum, Monad, HyperEVM and Solana.

Get a quote Read the FastLane reports
10+ Public staking reports
25+ Criticals and Highs in those reports
59 Invariants in the FastLane Protocol suite
10M+ Fuzzing runs per invariant in that suite

Smart contract audits for

  • GMX
  • USDT0
  • LayerZero
  • Ethena
  • Synthetix
  • Jupiter
  • Yuga Labs
  • Orderly Network
  • Olympus
  • Limit Break
  • M0
  • Dolomite

Onchain · Liquid Staking Audit

Every rate, every validator, every redemption.

  • Exchange rate accounting The rate between the LST and the underlying, held correct on deposits, queued withdrawals, rebases and every inherited ERC-4626 function.
  • Deposits and minting One-for-one issuance at the current rate, with no mint path that skips pricing, fees or a pending loss.
  • Withdrawal and redemption queues Queue order, idle liquidity gates, who can complete whose request, and entries that can jam the line.
  • Instant unstaking pools Atomic liquidity, utilization fees and caps that hold at full capacity and refill without leaks.
  • Validator and operator delegation Stake allocation across validators, adding and removing them, and cranks that must always complete.
  • Slashing and loss socialization How a loss reaches the rate, and whether holders, pending exits and new minters each carry their share.
  • Reward accrual and fee skims Rewards booked once, protocol and manager fees on the right base, and compounding nobody can sandwich.
  • Unbonding periods and epochs Epoch boundaries, unbonding delays and the bookkeeping that carries unstake demand from one epoch to the next.
  • EigenLayer restaking and LRTs Delegation, strategy shares and queued withdrawals, checked against the restaking release actually deployed.
  • Cross-layer balances HyperCore and HyperEVM balances, and funds in transit between them, counted once in backing.
  • MEV and priority fee capture Priority fees and MEV revenue routed to stakers, each booked to its real source rather than counted as staking yield.
  • Admin roles and pauses Who can mint, pause withdrawals, set fees, add operators or upgrade, with their thresholds and delays.

The LST audit covers liquid staking tokens, liquid restaking tokens, EigenLayer vaults and native staking pools on Ethereum, L2s, Monad, HyperEVM and Solana. Oracle reporters, rebase bots and the operator scripts that move stake are reviewed in an Offchain Automation Pentest.

Get a Liquid Staking Audit

Public findings

What Guardian has found in LST and restaking code.

Each card links a finding in a public Guardian report. All were resolved.

  1. C-01Orderly Solana Staking · Critical

    Solana reward claims were meant to pass through send_claim, which checks a merkle proof. The send_request program never rejected the claim payload type, so an attacker could name any amount and drain every ORDER token from the reward manager on Orderly’s Ledger chain.

    Reward claims Read the finding C-01 in the Orderly Solana Staking report

  2. H-01Valantis Overseer · High

    Mint counted the new deposit before checking for a pending slash, so a minter could plug the shortfall and mint stHYPE 1:1. The loss meant for existing holders then fell on the new minter.

    Slashing Read the finding H-01 in the Valantis Overseer report

  3. VLT-1Rest Finance · High

    EigenLayer cut the vault’s strategy shares when a withdrawal was queued, but restETH was burned only when it completed. In between, restETH’s price dropped.

    Exchange rate Read the finding VLT-1 in the Rest Finance EigenLayer LST Vault report

  4. VLT-2Rest Finance · High

    The vault never overrode ERC-4626’s mint function, which valued the flagship LST 1:1 with ETH. When the LST traded far enough above ETH to cover withdrawal fees, an attacker could mint through it and withdraw the difference from the vault.

    Minting Read the finding VLT-2 in the Rest Finance EigenLayer LST Vault report

  5. H-01Valantis Staking Modules · High

    Anyone could burn stHYPE to a contract that reverts on receiving HYPE. The stuck entry kept holding idle liquidity, so later burners could not redeem until governance found more HYPE.

    Redemption queue Read the finding H-01 in the Valantis Staking Modules report

  6. H-01FastLane Protocol · High

    When the atomic unstaking pool ran short, the fee math capped the gross amount but still sent the agent the full net amount, and the fee was never collected. Repeated each time the pool refilled, it drained instant liquidity fee-free.

    Instant unstaking Read the finding H-01 in the FastLane Protocol report

  7. H-03FastLane Protocol · High

    After an atomic unstake and a large unstake request, a target percentage in the global crank overflowed and reverted. With the crank stuck, requested unstakes could never complete.

    Unbonding Read the finding H-03 in the FastLane Protocol report

  8. M-04FastLane ShMonad Updates · Medium

    After a failed unwind, removing a validator deleted its records without checking for pooled stake still delegated. When the owner re-added the same ID, that stake was booked as new and counted twice in the total that prices shMON.

    Validators Read the finding M-04 in the FastLane ShMonad Updates report

  9. EWT-1Rest Finance · High

    Withdrawals called queueWithdrawals on EigenLayer’s DelegationManager, which the deployed M1 contracts did not support. Every attempt would revert, leaving users unable to withdraw restaked assets.

    Restaking Read the finding EWT-1 in the Rest Finance EigenLayer LST Vault report

  10. H-01Bracket LST Vault · High

    updateNav added cumulative manager fees to only the current epoch’s protocol fee. That total sets what the manager must transfer to the vault, so each update asked for the wrong amount.

    Fee skims Read the finding H-01 in the Bracket LST Vault, Round 2 report

  11. STK-1Key Finance · High

    Staking GMXkey had no fee or lockup, so an attacker could buy and stake just before rewards compounded, claim a share, then unstake and sell. Rewards meant for other stakers went to the sandwich.

    Reward accrual Read the finding STK-1 in the Key Finance GMX Key report

Onchain · How it runs

Two teams, and the backing written as code.

  1. 01 / Two teams

    Two independent teams, one scope.

    Team A pairs expert manual review with the fuzzing suite. Team B runs Helix across every corner. Each cross-checks the other.

    • Expert manual review
    • Competitive incentives
    • Cross-checked findings

  2. 02 / Helix AI Auditor

    Helix, the leading AI Auditor.

    Hundreds of parallel agents spend billions of tokens across the entire codebase. Only human-verified findings reach the report.

    • Billions of tokens
    • Human-guided context
    • Human-triaged findings

  3. 03 / Invariant fuzzing

    An invariant suite around your backing.

    Core safety properties become executable invariants, fuzzed across every reachable state. On a staking protocol, those are its backing and accounting rules.

    • No round-trip profit in one epoch
    • Balance always covers accrued fees
    • Tokens cover every obligation

  4. 04 / The staking layer

    Read against the chain you ship to.

    Contracts are checked against the staking layer they call, as it is deployed: EigenLayer’s live release, HyperCore transfers and Monad’s staking precompile, with its gas costs.

    • Restaking release compatibility
    • Funds in transit between layers
    • Precompile calls and gas

The suite stays with your code for every later change. FastLane’s report lists 59 invariants, each fuzzed 10M+ times, and Bracket’s LST vault suite ran 20M+ times per invariant.

Get a Liquid Staking Audit

Published work

Staking protocols Guardian has reviewed.

Incidents

When liquid staking went wrong for holders.

  1. $5MAnkr aBNBc

    An attacker compromised Ankr’s BSC deployer key, upgraded the aBNBc token and, per Rekt, minted 60 trillion aBNBc. Thin liquidity capped the take at about $5M.

    Dec 2022 Our coverage of the Ankr aBNBc exploit

  2. $26MAave wstETH

    A cap in Aave’s CAPO oracle priced wstETH about 2.85% below its exchange rate. Chaos Labs counted about $26M of liquidations across 34 accounts, and no bad debt.

    Mar 2026 Our coverage of the wstETH oracle liquidations

  3. 8%stETH market discount

    Before Ethereum allowed staking withdrawals, the only exit from stETH was a sale. CoinDesk reported the discount hit a record 8% on June 13, 2022, and by June 16 Curve’s pool held about 111,000 ETH against 492,000 stETH.

    Jun 2022 Our coverage of the 2022 stETH discount

“Guardian has been an excellent security partner through multiple audits. Their combination of great security researchers and custom fuzz testing is an incredibly strong approach to security. Highly recommend.”
Core TeamFastlane
“Guardian found a key bug that had been missed in previous audits, despite being outside the audit’s scope. This is an honest signal of commitment and going the extra mile is rare.”
EdValantis

Resource Hub

Before you book a staking smart contract audit.

Questions

Liquid staking audit questions

What does a liquid staking audit cover?

The contracts that mint, price, delegate, queue, slash and redeem, plus fee settings and admin roles, at one frozen commit. Scope counts nSLOC and lists each integration, such as the EigenLayer release or staking precompile you call. Our scoping guide lists what to send.

How is a restaking or liquid restaking token audit different?

It adds the restaking layer’s rules on top of the LST’s. On EigenLayer, queued withdrawals stay slashable for about 14 days, so an LRT’s queue, rate and loss handling must cover that window. Our restaking risks guide walks through each layer.

Which invariants do you fuzz on a staking protocol?

Accounting and backing rules. FastLane’s report lists “Asset delta equals Liability+Equity delta”, which held, and “Goodwill during crank should be zero with all active validators”, which failed under fuzzing. Rest Finance’s suite checks “Vault Supply = Sum of Actor Share Balances”.

Do you audit liquid staking outside Ethereum?

Yes. Guardian’s public reports cover FastLane’s shMONAD on Monad, Valantis’s stHYPE on HyperEVM and Orderly’s staking program on Solana. Each chain changes the logic: Valantis’s review found the module’s balance missed HYPE in transit to HyperCore for at least one block.

How long does a liquid staking audit take?

It depends on size and complexity. In Guardian’s public reports, two auditors reviewed Valantis’s staking modules in three days, five reviewed Rest Finance’s EigenLayer vault in about a week, and four spent 17 days on FastLane’s protocol.

Can an audit show our LST meets the SEC’s liquid staking conditions?

No, that is a legal reading for counsel. SEC Release 33-11412 ties its view to one-for-one issuance and redemption at any time, subject to unbonding. An audit tests the mint and withdrawal code behind those conditions. Our SEC guidance post maps each one to code.

What happens when we change the code after the audit?

The report covers the audited commit and the fixes Guardian reviewed. Valantis returned two months after its staking modules review for a separate review of its Overseer contract. Vanguard adds six months of update review, up to 2,500 onchain SLOC.

Smart Contract Audit

Other engagements

Start now

Get a quote for your liquid staking audit.

Get a quote