Stader

Tier 2

Security Analysis Report

TVL

$156.73M

Sections

7

report sections

Citations

547

evidence references

Version

v5

11/14/2026

5.3

Moderate Risk

Safety Score · out of 10

📋

Executive Summary

5.3/10

Executive Summary: Stader

Stader is a liquid staking protocol. Its main product, ETHx on Ethereum (about $133.0M), is backed by staked ETH and has a 7-day delay on core upgrades, which can be shorter than the 7 to 10-day redemption queue. The weak points sit around that core: a fast 3-of-6 Manager multisig, a three-node oracle committee, and no audit that could be tied to the code now deployed. HBARX on Hedera (about $23.7M) has no visible audit, controls or mechanics in the reviewed sources.

Safety Rating

Overall Score: 5.3/10 Dimension scores: Code security 4.5 · Economic security 4.5 (provisional) · Team transparency 7.1 · Governance 4.9 · Track record 6.7

Scores by dimension, points out of 10 (points)
Code security
4.5 points
Economic security
4.5 points
Team transparency
7.1 points
Governance
4.9 points
Track record
6.7 points

The rating reflects a conventional pooled staking token with a well-timelocked core and a weaker periphery. Evidence confidence is moderate, because mainnet ETHx market depth was not measured and no audit could be tied to the deployed code version.

Scores by Dimension

DimensionKey Finding
Code SecurityMany audits across products, but none could be tied to a specific deployed version, so live code was not compared with audited code.
Economic SecurityETHx is backed by real staked ETH, but exits can be suspended with no stated bound and the loss-cover fund has no stated size or user-loss rule.
Team TransparencyThe team is an identifiable organisation, but only one person is named and his leadership role is unconfirmed by an official source.
GovernanceCore ETHx upgrades need a 7-day timelock and a 6-of-9 multisig, but a 3-of-6 Manager multisig acts instantly and voting is off-chain and concentrated.
Track RecordOperating since late 2021 with one documented exploit (NearX on NEAR, 2022-08-16, about USD 800k); completed repayment is not shown, and incidents after the supplied material could not be searched because both web searches failed on a provider quota error.

What This Protocol Does

Stader lets users stake ETH or HBAR and receive a tradable receipt token (ETHx or HBARX). For ETHx the token's value against ETH rises as staking rewards accrue. Today almost all TVL sits in ETHx on Ethereum plus HBARX on Hedera. Earlier products on Polygon and BNB Chain are being wound down.

The protocol also includes the SD token. SD is Stader's governance token, and permissionless node operators must bond SD as collateral for each validator. The SD Utility Pool is a lending pool where SD is lent to node operators. The docs read do not fully explain the pool's purpose.

DefiLlama lists Fantom, Terra2, Binance and Near as inactive chains with no current TVL. Stader was deployed there earlier and those deployments are not assessed.

Key Strengths

  • ETHx core upgrade control (ETHx on Ethereum). Core upgrades need 6 of 9 community multisig signatures and a 7-day delay. The delay is confirmed on-chain at Ethereum block 26384083 (2026-11-10). This is a limit, not a guaranteed exit window: the 7-day notice can be shorter than the typical 7 to 10-day redemption queue if six community signers are compromised (low likelihood, hypothetical; see Key Risks).
  • Deepest audit review is on ETHx. Halborn, Sigma Prime, Code4rena and MixBytes reviewed ETHx between 2023 and 2024. Other products have separate, older reviews. Several audit copies are hosted by the protocol or by unverified hosts, which lowers confidence in their provenance but not the assessment.
  • Source-matched deployments. 64 of the 66 EVM deployments in the inventory are Sourcify runtime matches, and 2 have explorer-listed source (an explorer claim).
  • First-loss bond for permissionless ETHx operators. The operator's 4 ETH collateral absorbs slashing losses first. This applies to the permissionless pool only; permissioned operators post no bond.
  • Incident disclosure (NearX, an inactive product). Stader published a post-mortem and a make-whole plan for the 2022 exploit. Completed repayment is not shown.

Key Risks

  • 1. ETHx exits frozen through the fast Manager multisig (ETHx, up to about $133.0M; likelihood low). Three of six Manager signers can pause 11 listed contracts at once. Unpausing needs the 7-day timelock on 7 of those 11 listed contracts, while the two node registries can be unpaused by the Manager multisig directly. Safe mode has no stated time limit. The reports show no path for the freeze to take principal directly. The cross-analysis rates the likelihood as low, since it needs three compromised or colluding keys or a mistaken pause.
  • 2. ETHx exchange rate set by three keys and an instantly upgradable oracle (ETHx and third-party integrators; likelihood low; loss not bounded by the reports). A three-node committee reports the ETH backing that sets the ETHx rate. That figure is a committee report, not an independent read of validator balances. The ETHxPriceOracle proxy upgrades through the Manager Safe with no delay. The reports do not show which lenders or chains read that oracle. The likelihood is low because it needs a committee majority or three Manager keys.
  • 3. HBARX on Hedera is unassessed (HBARX, up to about $23.7M, about 15% of TVL; likelihood unknown). The reports show no mechanics, controls or audit for HBARX. The cross-analysis gives no likelihood band; the Hedera admin keys, deployed source and the linked Halborn audit would decide it.
  • Core upgrade notice shorter than the exit queue (ETHx; impact catastrophic, likelihood low; hypothesis). This scenario needs an attacker to compromise six of the nine Community multisig signers and schedule a harmful core upgrade. The 7-day notice can be shorter than the typical 7 to 10-day redemption queue, so some requests may not finish before the upgrade executes. The impact is catastrophic for the core ETHx pool (up to about $133.0M), and 7 days of notice does not guarantee that exits finish. The likelihood is low: six signatures plus public notice are substantial barriers, and no such event is on record.
  • Correlated slashing with an unbonded permissioned pool (ETHx; impact severe, likelihood low). A correlated slashing event could exceed the 4 ETH bond, and the sources do not show what absorbs permissioned-operator losses. The insurance fund's documented flows cover operator front-running only.

Critical Findings

Audit findings below cover all products and audits, not only ETHx. Counts are groups of findings, merged across audits, with the latest auditor status. Several audit copies are hosted by the protocol or by unverified hosts, which lowers confidence in their provenance but not the assessment.

SeverityTotal groupsFixed per the auditorNot marked fixed
Critical880
High31274
Medium714427

"Not marked fixed" follows the auditor's final status. It is not a count of confirmed live vulnerabilities.

Critical: no group is marked not fixed.

High, not marked fixed (4 groups, all MaticX on Polygon, a product being wound down):

  • Missing Authentication (Immunebytes, 2022-04-25). Status: acknowledged. Our code check is unavailable because the audited code could not be located.
  • Governance Privilege: contract owner has control over funds (Bail Security, 2024-09). High (estimated; the audit label was "Governance"). Status: acknowledged. Code check: the checked version does not contain the described code in this form; proxy administration and multisig ownership are outside the supplied file.
  • Governance of ValidatorShare and StakeManager contracts (Bail Security, 2024-09). High (estimated; the audit label was "Governance"). Status: acknowledged. Code check: the checked version does not contain the described code in this form; the Polygon functions cited are not in the supplied file.
  • MaticX conversion susceptible to an inflation attack after a new deployment (Bail Security, 2024-09). High (estimated; the audit label was "applicable only to forking"). Status: acknowledged. Code check: the default-branch snapshot may still contain the described code (low confidence; deployment not confirmed).

Medium, not marked fixed (27 groups):

ETHx and SD token on Ethereum (audited code versions were not compared with deployed code; "deployed code" means the source verified for the live contract):

  • Slashing of a validator can exceed the actual penalty due (Halborn, 2023-04-24). Status: partially fixed. Code check: the described code is unchanged in the verified SDCollateral source (Sourcify runtime match; identical to public commit stader-labs/ethx@1939e6c).
  • Protocol will not benefit from the slashing mechanism when the remaining penalty is bigger than minThreshold (Code4rena, 2023-06-02). Status: acknowledged. Code check: the described code is unchanged in the same verified SDCollateral source.
  • MEV bots can win all the auctions when Auction is paused (Code4rena). Status: disputed. Code check: a fix is visible in the verified Auction source, which shows no pausing base.
  • Owner in VaultProxy.sol is address(0) (Code4rena). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form.
  • ValidatorWithdrawalVault.distributeRewards can be called to make an operator slashable (Code4rena). Status: disputed. Code check: disputed by the team; the described code is unchanged in the verified vault source; not scored (deployed code, low confidence).
  • ValidatorWithdrawalVault.settleFunds does not check the amount a user has inside NodeELRewardVault to pay a penalty (Code4rena). Status: acknowledged. Code check: the described code is unchanged in the verified vault source, and the same holds for the node reward vault (deployed code, low confidence).
  • No bidder has an incentive to bid in the Auction except last-minute MEV, due to a fixed endBlock (Code4rena). Status: acknowledged. Code check: the described code is unchanged in the verified Auction source (deployed code).
  • Chainlink latestRoundData may return a stale or incorrect result (Code4rena 2023-06-02; also reported twice by Halborn, 2023-04-24). Status: acknowledged. Code check: the described code is unchanged in the verified StaderOracle source (deployed code).
  • HasEnoughSDCollateral() check is performed only once during onboarding (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form (deployed code).
  • Submit functions are susceptible to front-running when trusted nodes are removed (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the described code is unchanged in the verified StaderOracle source (deployed code).
  • Uninitialized WithdrawDelay (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the audited claimWithdraw and withdrawDelay are absent from the deployed code, which uses a different withdrawal flow.
  • Lack of pause functionality within the OFT Adapter (MixBytes, 2024-07-11). Status: acknowledged. Code check: the checked version does not contain the described code in this form, but the inherited base contract is not among the supplied files, so an inherited pause cannot be ruled out.

MaticX on Polygon (checked only against the default-branch snapshot stader-labs/maticX@b27e328; deployment not confirmed):

  • Reentrancy guard missing in MaticX.requestWithdraw (Immunebytes, 2022-04-25). Status: acknowledged. Code check: a fix is visible in the snapshot (nonReentrant on the withdrawal request).
  • A malicious user can block withdrawals by dusting the contract with small ValidatorShare token amounts (Bail Security, 2024-09). Status: acknowledged. Code check: the snapshot may still contain the described code (low confidence).
  • Lack of reward compounding before submit lets users extract value (Bail Security, 2024-09). Status: acknowledged. Code check: the checked version does not contain the described code in this form.
  • Exotic edge case will result in a bricked submit function (Bail Security, 2024-09). Medium (estimated: the audit label was “applicable only to forking”, and the case can lock deposits). Status: acknowledged. Code check: the snapshot may still contain the described code (low confidence).

BnbX on BNB Chain (checked only against the default-branch snapshot stader-labs/bnbX@18442f8; deployment not confirmed):

  • Governance privilege: users are completely dependent on the operator to withdraw funds (Bail Security, 2024-07). Medium (estimated; the audit label was “Governance”). Status: acknowledged. Code check: the snapshot may still contain the described code (low confidence); starting a batch undelegation remains operator-only.
  • Lack of exchange-rate update allows users to flash-theft tokens (Bail Security, 2024-07). Status: acknowledged. Code check: the checked version does not contain the described code in this form; the recommended change is only partly visible.

NearX and a December 2021 report (inactive chains):

  • Potential Centralization Problem (I) and (II) (BlockSec, NearX, 2022-10-13). Medium (estimated: the audit gave no severity). Status: acknowledged. Code check: unavailable, because the audited repository is not available.
  • Possibility of for loop exceeds block (December 2021 report by The Z Institute; weak source). Status: acknowledged. Code check: the checked version does not contain the described code in this form (NearX snapshot stader-labs/nearx@3912ee1).

Some rows above hold several findings, so the rows are not a one-to-one count of the 27 groups. The slashing rows are the clearest recurring pattern. The audits do not describe either as a direct drain of user funds. “Described code still there” is a limited code signal, not a confirmed vulnerability.

Other material issues from the reports:

  • Audit-to-deployment gap (all EVM products): For all 33 deployed components the audited version was not identified. This is a comparison that was not performed, not a measured mismatch.
  • HBARX on Hedera has no audit report in the evidence. The linked Halborn audit could not be retrieved, so its date, scope and applicability are unknown.
  • For six Sigma Prime ETHx findings marked fixed, the deployed files differ from the fix revision the auditor reviewed. Those fixes cannot be confirmed as deployed as reviewed.
  • Documented versus on-chain: an older official docs page says the Manager multisig is a 3-of-5. On-chain on Ethereum at block 26384083 (2026-11-10) it is a Safe with threshold 3 and 6 owners.
  • Documented versus on-chain: the docs describe the 7-day timelock as owning all ETHx contracts. On-chain at block 26384083 the ETHxPriceOracle and the ETHx proxies on Arbitrum, Optimism and Gnosis upgrade through 3-signature Safes with no timelock in the path.
  • Documented versus on-chain: the docs promise at most one oracle operator change per week. The on-chain cooling period for trusted-node changes reads zero at block 26384083, so the limit is not visible in that setting. Another mechanism may exist.
  • Documented versus on-chain: the ETHxRateProvider stores a rate of 1.0335 last updated on 2024-06-28, while its live read at block 26384083 returns 1.1014. Which value other chains see is not confirmed.
  • The PermissionlessNodeRegistry on Ethereum read as paused at block 26384083. The reason is not stated.

Recommendation

Suitable for risk-tolerant users who want ETHx staking exposure and accept multisig-run controls. ETHx has a 7-day timelock on core upgrades and deep audit review, but the timelock is not a guaranteed exit window (a low-likelihood, hypothetical six-signer compromise could leave redemptions unfinished), and fast Manager keys, a three-node oracle committee, an unquantified loss fund and unmeasured mainnet exit liquidity remain. Users with low risk tolerance should wait for the Manager signers to be identified and for audited versions to be matched to deployed code. Treat HBARX as unassessed until its audit, controls and mechanics are published. TVL also fell between 2026-10-26 and 2026-11-13, and the sources do not explain the drop.

Report Sections

  • Protocol Overview
  • Code Security
  • Economic Security
  • Social & Team
  • Governance
  • Cross-Risk Analysis

94 citations · click markers to view evidence

🏗

Protocol Overview

Protocol Overview: Stader

Stader is a liquid staking protocol: users stake ETH or HBAR and receive a tradable receipt token (ETHx or HBARX). For ETHx the token's value against ETH rises as staking rewards accrue; the supplied sources do not confirm HBARX's exchange-rate method. Today its TVL sits almost entirely in ETHx on Ethereum (about $133M) plus HBARX on Hedera (about $24M). Earlier products on Polygon and BNB Chain are being wound down, and the NEAR product suffered an exploit in 2022. The ETHx mechanics and on-chain controls are well documented; HBARX mechanics, controls and audits are not.

What It Does

Stader lets users stake crypto without locking it up. A user deposits ETH and receives ETHx, a token that stays tradable while the deposited ETH earns Ethereum validator rewards. The ETHx price against ETH rises over time instead of the user receiving new tokens. Stader also offers HBARX on Hedera as a staking receipt; see the Hedera paragraph under How It Works.

  • Category: liquid staking.
  • Product set today: ETHx on Ethereum and HBARX on Hedera. The official docs list both as live networks and mark Polygon (MaticX) and BNB Chain (BNBx) as discontinued.
  • Other token: SD is Stader's governance token. Permissionless node operators must bond SD as collateral for each validator.
  • Inactive chains: DeFiLlama lists Fantom, Terra2, Binance and Near as inactive chains with no current TVL. Stader was deployed there earlier and those deployments are not assessed here.
  • Users: ETH holders who want staking yield plus DeFi use of the receipt token; home stakers and professional operators who run validators; SD holders who take part in governance.

How It Works

ETHx works as a pool: stakers supply the ETH, node operators run the validators, and an oracle reports rewards so the ETHx exchange rate can rise.

User flow for ETHx on Ethereum:

  1. The user deposits ETH into the staking contracts and receives ETHx. The deposit contract is StaderStakingPoolManager, which holds the ETHx minter role.
  2. The manager splits the ETH between a permissionless pool and a permissioned pool of node operators.
  3. Each operator runs a validator, and Stader lends the operator the ETH needed to fill it. Permissionless operators put up 4 ETH plus 0.4 ETH worth of SD as collateral. Permissioned operators are onboarded after DAO approval and can run validators without a security bond.
  4. Validator rewards accrue to the pool. The ETHx/ETH exchange rate is total ETH in the pool divided by ETHx in circulation, so it rises as rewards arrive.
  5. To exit, the user requests unstaking and later claims ETH. The payout uses the lower of the exchange rate at request time and at finalization. Requests typically take 7 to 10 days, subject to Ethereum entry and exit queues.
  6. The UserWithdrawalManager holds the burner role on ETHx for these redemptions.

Core mechanism: pooled staking with a rising-rate receipt token. The permissionless pool uses a 4 ETH bond per validator, which the litepaper calls the lowest in the ecosystem.

Node-operator safeguards stated in the official ETHx security explainer. The explainer describes the 4 ETH collateral without separating the two pools. Per the node-operator docs, permissioned operators post no bond, so the sources do not show what absorbs losses caused by a permissioned operator:

  • The protocol checks withdrawal credentials before lending 28 ETH, and imposes a 3 ETH penalty if it detects front-running.
  • A 1 ETH penalty applies to MEV misappropriation. MEV is extra value validators can capture from block ordering.
  • The operator's 4 ETH collateral absorbs slashing losses first. Slashing is the network's penalty for validator misbehaviour.
  • A safe mode disables withdrawals when many validators are slashed or more than 50% of ETHx validators are down.

Key Ethereum contracts, from the official contract list:

ContractRole (as documented or measured)
ETHxReceipt token; mint and burn restricted to the pool manager and withdrawal manager
StaderStakingPoolManagerAccepts deposits, mints ETHx
UserWithdrawalManagerHandles unstake requests and claims
PermissionedPool / PermissionlessPoolRoute ETH to the two operator sets
PermissionedNodeRegistry / PermissionlessNodeRegistryOperator and validator registries
StaderOracleReports rewards and exchange rate (see Dependencies)
StaderConfigCentral address and parameter registry
SDCollateral, SDUtilityPoolSD bonding and SD lending/borrowing for operators
Penalty, OperatorRewardCollector, SocializingPoolPenalties, operator reward payout, reward sharing
ETHx deposit, reward and unstake flow on EthereumOpen full view →
ACTORETH stakerCONTRACTStaderStakingPoolMana…TOKENETHxCONTRACTPermissioned andpermissionless poolsCONTRACTUserWithdrawalManagerdeposits ETH into1mints1splits ETH across2requests unstakevia5burns6
  • (actor)
  • (contract)
  • (token)
  • (contract)
  • (contract)
  • step 1flow
  • step 1flow
  • step 2flow
  • step 5flow
  • step 6flow

Hedera (HBARX). The official docs describe HBARX as the way to "Stake to earn staking rewards with HBARx and access DeFi opportunities". A Stader gitbook page (source reliability: unverified publisher) lists four HBARX mainnet contracts by Hedera ID: token 0.0.834116, staking 0.0.1412503, rewards 0.0.1412524 and undelegation 0.0.1412465. These addresses do not show how the contracts work, whether they are upgradeable, or who controls them. The page also links to HBARX documentation and security pages that were not retrieved, so the exchange-rate method, withdrawal path, controls and audits remain unconfirmed.

SD Utility Pool. The SDUtilityPool stores risk settings such as loan-to-value and liquidation threshold, and tracks delegator withdrawals and operator liquidations. Assessment: it works as an SD lending market for operators' collateral needs; the docs read here do not fully explain its purpose.

Wind-down products. Per Stader's gitbook pages (source reliability: unverified publisher), MaticX stopped taking deposits on 13 June 2026, its app was disabled on 3 August 2026, and contract withdrawals run until 3 August 2029. BNBx deposits were disabled on 29 January 2026 and its app on 9 June 2026. The official site lists both products as discontinued.

Yield Sources

ETHx yield comes from Ethereum validator rewards. The docs deduct 10% of those rewards and split it between Stader's protocol fee and node operators.

  • Source: the docs say ETHx rewards "depend on the rewards released on both the beacon and consensus layers". Validators earn consensus rewards, tips and MEV.
  • Fee: 10% of rewards is deducted, split 5% to the protocol and 5% to node operators. On-chain, StaderConfig returns a raw getTotalFee of 10000 at block 26384083; the units were not verified, so this is not confirmed as 10%.
  • Historical rate: the docs quote "approximately 3.5%" APR, but the page carries no date in the text read, so treat it as indicative only.
  • Extra incentive: permissionless operators earn SD rewards on bonded SD collateral. This is a token incentive paid to operators, not to stakers.
  • Protocol revenue: DeFiLlama records fees of 352,300 USD over 30 days and 7,122,874 USD over 365 days, and revenue of 20,999 USD (30 days) and 383,878 USD (365 days). DeFiLlama's definitions of "fees" and "revenue" for Stader are not in the sources.
  • Sustainability: Assessment: ETHx yield is sustainable in the sense that it rests on Ethereum consensus rewards. It rises and falls with network issuance and MEV. HBARX yield sources are not documented in the supplied sources.

Dependencies & Integrations

ETHx depends on Ethereum's beacon chain, Stader's own oracle network, and (for other chains) LayerZero messaging.

  • Ethereum staking: validators deposit into Ethereum's official deposit contract, which StaderConfig references on-chain.
  • Stader oracle: the StaderOracle reports exchange rates and validator data. The docs say oracle operators post security collateral, and the oracle needs "a strict majority of Oracle operators to function". Large rate jumps trigger an inspection mode: "Critical updates like Exchange Rates have built-in guardrails, enabling inspection mode if two subsequent rates deviate significantly." Oracle collateral sits in a 3-of-5 multisig.
  • Cross-chain rate and token: the ETHx rate oracle repository is "Based on gyarados rate oracle using LayerZero v1". On-chain reads show an ETHxRateProvider and an ETHx_OFTAdapter on Ethereum, both pointing at a LayerZero endpoint, and an ETHx_OFT plus LayerZero endpoint on Gnosis. The bridge is LayerZero-based; its verifier settings were not read.
  • Other dependency listed: DeFiLlama's dependency map lists Chainlink CCIP as a bridge dependency with share: 0.0, and marks it unresearched. How Stader uses CCIP is not established.
  • DeFi use: the docs name yield farming, lending, borrowing and liquidity pools as uses for ETHx. The litepaper (pre-launch) named Aave v3, Balancer, Beefy and others for Stader's earlier tokens; current integrations were not confirmed.
  • Chains: DeFiLlama lists active TVL on Ethereum and Hedera.
  • Polygon dependency (wind-down): Stader's gitbook page (source reliability: unverified publisher) says Polygon-address holders can either use the Polygon-side flow or move MaticX to Ethereum with the Polygon native bridge.

TVL & Adoption

Stader's TVL is $156.7M on the latest DeFiLlama capture (13 November 2026), down sharply over the previous month.

  • Latest snapshot: , collected 13 November 2026.
  • Chain split: Ethereum $133.0M and Hedera $23.7M; the other four chains hold zero.
  • Trend: DeFiLlama's daily series shows 207.7M USD on 26 October 2026 and 162.9M USD on 13 November 2026. Over 30 days TVL fell about $53.0M, of which about $41.8M is a price effect and the rest is a fall in staked quantity at current prices.
  • Peak: TVL in the series was 869.2M USD on 1 April 2022, mostly from earlier multi-chain products.
  • ETHx supply: the ETHx token's total supply read on-chain is a raw 62381254843589321153940 at 18 decimals, about 62,381 ETHx.
  • Token market cap: DeFiLlama shows an SD market cap of about $7.35M.
TVL by chain in USD millions, DefiLlama snapshot 13 Nov 2026 (USD m)
Ethereum
133.0 USD m
Hedera
23.7 USD m

Technical Architecture

ETHx contracts are upgradeable proxies owned by a 7-day timelock, with a fast manager multisig that can pause. Measured on Ethereum at block 26384083 (state read 10 November 2026). For the full control analysis see Governance.

  • Upgrade pattern: most ETHx contracts are EIP-1967 proxies (a proxy forwards calls to replaceable logic code). One shared proxy admin owns upgrades, and its owner is the timelock at 0x1112…68B9. The fastest upgrade delay is 604800 seconds, or 7 days.
  • Timelock: minimum delay 604800 seconds. The community multisig is both proposer and canceller; the executor role is held by the zero address, which in this timelock style means anyone may execute after the delay.
  • Community multisig: the official governance page describes a 6-of-9 group that can "unpause and upgrade contracts via a time lock set with a minimum delay of 7 days" and rotate the Manager and Operator addresses. The Safe at 0x45B9…9fF6 reads as threshold 6 of 9 owners on-chain.
  • Manager multisig: the official governance page says it is a 3-of-5 that can pause and change configuration. The Safe at 0xAAfb…07AF reads as threshold 3 of 6 on-chain at block 26384083, so the governance page's owner count differs from the on-chain state. It holds the manager role and an operator role in StaderConfig; the timelock holds the config admin role.
  • Operator role: the docs say the Stader team controls the Operator role. On-chain at block 26384083 the operator role has two holders: the manager multisig and 0x8f30…737e, which has no deployed code and so is a single key (EOA).
  • Pause asymmetry: the manager multisig can pause ETHx, the pool manager, the withdrawal manager and the oracle with no delay. Unpausing those contracts goes through the 7-day timelock. The two node registries can be paused and unpaused by the manager multisig directly.
  • Current pause state: the Ethereum PermissionlessNodeRegistry (0x4f4B…cF1) read as paused at block 26384083, while ETHx, the pool manager and the withdrawal manager read as not paused. The reason is not stated in the sources. Assessment: new permissionless validator onboarding is likely halted; effects on deposits and withdrawals are not documented.
  • Parameters changeable by governance or admin: withdrawal limits and block-delay settings in StaderConfig (timelock-gated, 7 days), safe mode, oracle settings and fees.
  • Emergency mechanisms: pause, safe mode and pre-signed validator exits that Stader stores and can broadcast.
  • Treasury and SD rewards: the Stader treasury address reads as a Safe with threshold 2 of 5 owners.
  • Not read: four contracts on Arbitrum Sepolia (chain 421614) were not read on-chain.

Audit coverage (ETHx only; other products are separate):

  • Stader's official security explainer says ETHx has been audited by SigmaPrime, Halborn, and Code4rena.
  • A MixBytes ETHx audit report is dated 11 July 2024. The audit index also lists a Sigma Prime report on the SD Utility Pool (date not stated in the index entry read), which covers that component only.
  • Audit-to-deployed-version comparisons were not performed. The coverage map lists 31 components as "in scope, version unknown" and 39 as without an attributed address.
  • Several audit files have unverified publication hosts; they still count as evidence at lower confidence.
  • No audit of the Hedera HBARX contracts appears in the audit index.
  • The $1M Immunefi bug bounty was described as "launching"; its current status was not confirmed.

Key Risks (Protocol-Level)

  • Where funds sit: staker ETH sits in the pool and registry contracts and in validators on Ethereum's beacon chain, not in a user wallet. Slashing or poor performance can reduce pool value; for permissionless operators the 4 ETH bond absorbs the first losses, while permissioned operators post no bond.
  • Admin control: the manager multisig can pause instantly, and a 7-day timelock gates upgrades. A compromised community multisig could schedule a malicious upgrade, and the 7-day delay is the only on-chain warning window. Likelihood: low; impact: high.
  • Oracle risk: the exchange rate depends on Stader's oracle operators. A wrong rate update could misprice ETHx; the docs cite majority consensus and an inspection-mode guardrail, which the sources do not show being tested.
  • Cross-chain risk: ETHx on other chains depends on a LayerZero-based OFT and a rate provider. Bridge security settings were not read.
  • Past exploit: Stader's own account says an exploit was found and used on its NEAR liquid staking product on 16 August 2022, caused by an edge case in Stader's NearX token implementation, with loss limited to 165k NEAR (about 800k USD). A third-party hack-tracking page (source reliability: unverified publisher) adds that the attacker minted 20 million NearX without matching staked NEAR and swapped it in NEAR pools; Stader's account does not confirm those details. That product is separate from ETHx and HBARX; the incident does not show a flaw in ETHx code.
  • Complexity: ETHx has about 20 core contracts plus oracle, node software and a lending pool for SD; the SD Utility Pool adds liquidation logic.
  • Hedera: about 15% of TVL sits in HBARX, with contract IDs listed but no mechanics, controls or audit visible in the sources. Insufficient data for a risk assessment; the HBARX documentation and security pages named by Stader could not be retrieved.
  • Concentration and shrinking TVL: TVL fell from 207.7M to 162.9M USD between 26 October and 13 November 2026 (DeFiLlama series). The sources do not explain the drop.
  • Design assumptions: that a 4 ETH bond covers slashing tail risk (Stader's own simulation with SSV), and that operators keep enough SD collateral.

Data Completeness

  • Available: DeFiLlama TVL and history (13 November 2026), ETHx docs and litepaper, official multisig descriptions, on-chain control reads for Ethereum at block 26384083 (10 November 2026), incident reports, audit index.
  • Missing: HBARX mechanics, contract code and controls, and audits (only a list of contract IDs from an unverified publisher is available; Stader's HBARX documentation and security pages could not be retrieved); the cause of the paused PermissionlessNodeRegistry; LayerZero verifier settings; current APR and DeFi integrations; current status of the bug bounty; the deployed fee denominator.
  • Not compared: audited code versus deployed implementations.
  • Limits: the litepaper is pre-launch marketing; FAQ pages are undated or older. No Blockworks reports were in the workspace.
  • Open: whether the 2022 NearX holders were fully reimbursed. A reimbursement plan exists but its completion was not confirmed.

101 citations · click markers to view evidence

🔒

Code Security

4.5/10

Code Security Analysis: Stader

Stader has a broad audit record: reviews by several named firms across ETHx, MaticX, BnbX, NearX and the retired Terra product, plus a USD 1,000,000 Immunefi bounty and verified or explorer-listed source for the 66 EVM deployments in the verification inventory. The weak point is traceability: none of the audits could be tied to a specific deployed version, so the code now live was not compared with what the auditors reviewed. Of 110 critical, high and medium issue groups, 79 are marked fixed by the auditors, none of the 8 critical groups is open, and 31 are not marked fixed, mostly acknowledged design or governance points. Hedera, which holds about USD 23.7M of the roughly USD 156.7M of DeFiLlama TVL, has no audit report in the evidence; the Hedera page links a Halborn audit that could not be retrieved.

Data Completeness

  • Issue counts below are groups of critical, high and medium findings, merged across audits and counted once with the latest auditor status. They are not counts of audit entries, and low and informational findings are not counted.
  • The review covered 26 audit sources, two of which (the protocol's audit list page and the ETHx litepaper) are not audit reports. I read audit sections to quote and check individual findings, not every audit in full.
  • The audited versions of the code could not be identified, so deployed code was not compared with audited code. This is a gap in what could be shown.
  • Not published or not collected: 12 audited code revisions and 3 repositories (nearx-exchange-rate-feed, nearx-aurora, stader-protocol-v0) are not published; market depth is not published; address attribution produced no valid result; and the source for one Ethereum contract used for the control read (0x3517d4c4…) is not verified.
  • 39 audit-named contracts have no attributed current address (for example MaticX, ValidatorShare, StakeManager, VaultProxy, StakeManagerV2), so their deployed code is neither verified nor compared. The Hedera (HBARX) contracts are also outside the verified-source inventory.
  • DeFiLlama lists TVL on Ethereum (USD 133.0M) and Hedera . Fantom, Terra2, Binance and Near are listed as inactive and are not assessed. Hedera hosts the HBARX product (contract IDs 0.0.834116, 0.0.1412503, 0.0.1412524 and 0.0.1412465 ); no audit report in the evidence names these contracts and no Hedera contract source is in the evidence. Stader's Hedera page links a Halborn audit titled "Hedera Stader Protocol v3 Smart Contract Security Audit" . Two attempts to retrieve that report failed on a provider quota limit, so its date, scope, findings and applicability to the current Hedera contracts are not known; the link alone is not a report. The documentation says MaticX and BNBx are being sunset, so their audits matter mainly for remaining claims.
  • Several audits are hosted by the protocol or by unverified hosts, which lowers confidence in their provenance but not the assessment.

Audit Coverage

Audits cover products separately; a review of one product does not cover another.

ProductAuditorDateScope
ETHxHalbornengagement April 24 to July 4, 2023ETHx contracts
ETHxSigma Prime ("ETHx Phase 2")2023-06-01ETHx contracts
ETHxCode4rena (public contest)2023-06-02ETHx contracts
ETHxMixBytes2024-07-11ETHx.sol and the cross-chain OFT adapter
ETHx (off-chain)Halborn2023-06-01 (oracle, Go), 2023-06-01 (permissionless node, Go), 2023-07-05 (pentest)off-chain software
ETHx (off-chain)Sigma Prime2023-07permissioned node command-line tool
SD tokenSigma Prime2024-01SD Utility Pool
SD tokenHalbornengagement July 26 to August 4, 2022SD staking ERC20
MaticXImmunebytes2022-04-25MaticX contracts
MaticXHalbornengagement April 3 to 26, 2022MaticX contracts
MaticXBail Security2024-09MaticX (including new deployment)
BnbXHalborn2022-07-13BnbX contracts
BnbXPeckShield2022-07-24StakeManager
BnbXBail Security2024-07BnbX
BnbXHalborn2024-07-10BNB migration
NearXHalborn (re-audit)2022-09-16NearX contracts
NearXBlockSec2022-10-13NearX exchange rate feed and Aurora staking
LunaX (Terra, retired)Oak Security2021-12-06Terra contracts
LunaX (Terra, retired)Halborn2022-01-17; 2022-03-09 (SD token staking, CosmWasm)Terra contracts
ERC20 tokenHalborn2022-01token contract
TerraThe Z Institute2021-12-05Terra staking, rewards and airdrops registry contracts

The December 2021 report names The Z Institute in its disclaimer ; the copy comes from an unverified publication host, so it is a weak source. Some other copies are first-party hosted (see Data Completeness). ETHx evidence:

Bug bounty: Stader's Immunefi program lists a maximum bounty of USD 1,000,000 . The page shows the program live since 08 July 2023 and last updated 01 January 2025, so its latest update is almost two years old as of this report. The evidence does not show a separate bounty for MaticX, BnbX or Hedera. For ETHx, the documentation lists active health alerts, including roles and permissions changes on deployed contracts and privileged address monitoring , and states that the contracts are pausable . These assurance measures are documented for ETHx only.

SeverityTotal groupsFixed per the auditorNot marked fixed
Critical880
High31274
Medium714427

Counts cover all products and audits. "Not marked fixed" follows the auditor's final status and is not a count of confirmed live vulnerabilities.

Deployed code vs audits

Summary: for all 33 deployed components the audited version was not identified, so 100.0% of the 14,747 lines in 90 verified files are in the "version not compared" bucket, and deployed code was not compared with the audited version. Source verification itself is strong.

Of 33 components, 31 are named in an audit scope with the version not identified, and 2 have an unconfirmed audit scope. Of 66 known deployments, 31 were compared with a public commit, 0 are identical to an audited revision, 0 show unaudited changes, and 4 searches are unresolved. By reviewed criticality, 13,526 lines are critical, 832 high and 389 peripheral, and all files are rated. For six Sigma Prime ETHx findings that the auditor marked fixed, the deployed files differ from the fix revision the auditor reviewed, so those fixes cannot be confirmed as deployed as reviewed; the other fixed findings have no fix review to compare. The table below lists the two unconfirmed-scope rows and the three unresolved rows first, then the rest.

ComponentDeployment; verificationAudit scopePublic code comparedCriticality
ProxyAdmin (ETHxPoolV1 ProxyAdmin)eip155:42161:0xe9609dcbc830784b3b68d5b4162c5193df5dfa98; Etherscan-listed source (explorer claim)audit scope not confirmed; matched to an audit scope by loose name onlynot compared: repository search incompletecritical
TimelockController (ETHx SecurityTech Explainer)eip155:1:0x1112d5c55670cb5144bf36114c20a122908068b9; Sourcify runtime matchaudit scope not confirmed; deployed source is a flattened file that cannot be comparednot compared: source representation unsupported
ETHx (implementation)eip155:42161:0xc54b43eaf921a5194c7973a4d65e055e5a1453c2; Sourcify runtime matchnamed in an audit scope; audited version not identifiednot compared: repository search incompletecritical
GnosisSafe (Community multisig contract address)eip155:1:0xd9db270c1b5e3bd161e8c8503c55ceabee709552 (implementation of proxy eip155:1:0x45b977cecb9dfaa17dfcba88826ef684b8489ff6); Sourcify runtime matchnamed in an audit scope; audited version not identifiednot compared: repository search incompletecritical
Stader (SD token address (Mainnet))eip155:1:0x30d20208d987713f46dfd34ef128bb16c404d10f; Sourcify runtime matchnamed in an audit scope; audited version not identifiednot compared: repository search incompletecritical
Auctioneip155:1:0xa2aa24e91a345b3dd8652f304390203d0e1c4d31 (implementation of proxy eip155:1:0x85a22763f94d703d2ee39e9374616ae4c1612569); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
ETHxeip155:100:0xb30256ca8a9ebe058eb78a4edbf3364e7f8e5d86 (implementation of proxy eip155:100:0x01af04690d17dc27b891a7f67e9eee4d14de8ea8) (+1 more); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@709c5c0; not compared with an audited versioncritical
ETHxeip155:10:0xbe23e1a64969cb28efdb6c3d2ce9e4bf16042187 (implementation of proxy eip155:10:0xc54b43eaf921a5194c7973a4d65e055e5a1453c2); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@09de94d; not compared with an audited versioncritical
ETHx (implementation)eip155:1:0x4c22ffd479637ea0ed61d451cbe6355627283358; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
ETHxPoolV1eip155:42161:0x463da194e0653d9f4bef5d342f18679d7404b6c4 (implementation of proxy eip155:42161:0x107e427848937c4217c0efc1c3f35091de08e45e); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@709c5c0; not compared with an audited versioncritical
ETHxPriceOracleeip155:1:0x6904603c27392310d19e389105ca792fb935c43c (implementation of proxy eip155:1:0x9305d41ac91a200b848d285242d8e415e6414b2f); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@709c5c0; not compared with an audited versionhigh
ETHxRateProvidereip155:1:0x0b2faadfe75fe6507dc8f81f8331198c65ca2c24; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@709c5c0; not compared with an audited versioncritical
ETHx_OFTeip155:10:0x01af04690d17dc27b891a7f67e9eee4d14de8ea8 (+2 more); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@709c5c0; not compared with an audited versioncritical
ETHx_OFTAdaptereip155:1:0x39d0ce7d1509a4f22619d0f666978fbe820665e9; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx_oft@6a4ffad; not compared with an audited versioncritical
NodeELRewardVault (NODE_EL_REWARD_VAULT)eip155:1:0x97c92752dd8a8947ce453d3e35d2cad5857367af; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@4e63182; not compared with an audited versioncritical
OperatorRewardsCollector (implementation)eip155:1:0x134fee660a3259d8885f29c37fb1a6980a211c3f; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
OperatorRewardsCollector (OperatorRewardCollector)eip155:1:0x922805c0097547663e0df537ac65bd9b65cb2a11 (implementation of proxy eip155:1:0x84ffdc9de310144d889540a49052f6d1adb2c335); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@9d4a921; not compared with an audited versioncritical
Penaltyeip155:1:0x8d80ad529f68b11c2492301ba4fb78e5a654c206 (implementation of proxy eip155:1:0x84645f1b80475992df2c65c28be6688d15dc6ed6); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
PermissionedNodeRegistryeip155:1:0x549cfb2ebbb21c20e1c35bf3b944cc76d13177b9 (implementation of proxy eip155:1:0xaf42d795a6d279e9dcc19dc0ee1ce3ecd4ecf5dd); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@9d4a921; not compared with an audited versioncritical
PermissionedNodeRegistry (implementation)eip155:1:0xf2b656b2fc829ac1450cad03dc302d680ebbc214; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
PermissionedPooleip155:1:0xf393d241258164e9779972447284fc4d47c3b4d1 (implementation of proxy eip155:1:0x09134c643a6b95d342bdaf081fa473338f066572); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
PermissionlessNodeRegistry (implementation)eip155:1:0x0addf12ee7d096e2a992046e3234d9b7a3788524; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
PermissionlessNodeRegistryeip155:1:0xc2c4844b5b27a2395c2f16a07ac7bb2e7e0f07ad (implementation of proxy eip155:1:0x4f4bfa0861f62309934a5551e0b2541ee82fdcf1); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@9d4a921; not compared with an audited versioncritical
PermissionlessPooleip155:1:0x30c2501b3b2031fba3000daa6f8ed0a42fbbb3f0 (implementation of proxy eip155:1:0xd1a72bd052e0d65b7c26d3dd97a98b74acbbb6c5); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
PoolSelectoreip155:1:0x8322793ceb2d8740ad83ed431e69ae87fa1a86ab (implementation of proxy eip155:1:0x62e0b431990ea128fe685e764fb04e7d604603b0); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@21d516c; not compared with an audited versioncritical
PoolUtilseip155:1:0xbaa0c93e3e24e863e86c7b508fc0dc57bc5718a0 (implementation of proxy eip155:1:0xeda89ed8f89d786d816f8e14cf8d2f90c6bf763f); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
SDCollateral (ETHx Mainnet Smart Contracts)eip155:1:0x278097f753ab3e4a5e897dbb72cea1592e78e9a5 (implementation of proxy eip155:1:0x7af4730cc8ebad1a050dcad5c03c33d2793ee91f); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
SocializingPool (ETHx Mainnet Smart Contracts)eip155:1:0xcc0dab27c23599cd7d2b7a04edf0730f8b2509a7 (implementation of proxy eip155:1:0x9d4c3166c59412cedbe7d901f5fde41903a1d6fc); Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
StaderConfig (implementation)eip155:1:0x2dba235892dc72efa74e0b395294494b5b0b32b2; Sourcify runtime matchnamed in an audit scope; audited version not identifiedpublic commit stader-labs/ethx@1939e6c; not compared with an audited versioncritical
Audit-named contracts with no attributed address (39 names, e.g. MaticX, StakeManager, StakeManagerV2, ValidatorShare, VaultProxy)nonenamed in audit scopes; no current address attributednot comparednot assessed (unknown is not low risk)

Showing 29 components and one grouped row; 4 further components with the same status are not shown. The inventory is partial: the 39 audit-named contract names have no attributed address, and unknown is not treated as low risk.

Findings by Severity

Only groups the auditor did not mark fixed are listed. Fixed groups are counts only. Ratings marked "(estimated)" are AI estimates, given where the audit gave a label such as "Governance" instead of a severity.

Critical

Total groups: 8; fixed per the auditor: 8; not marked fixed: 0. Nothing to list.

High

Total groups: 31; fixed per the auditor: 27; not marked fixed: 4.

  • Missing Authentication (Immunebytes, MaticX, 2022-04-25). Status: acknowledged. Code check: not marked fixed per the auditor (acknowledged); our code check is unavailable: the audited code could not be located.
  • Governance Privilege: Contract owner has control over funds (Bail Security, MaticX, 2024-09). High (estimated: the audit label was "Governance", and the finding says the proxy admin has full control over user funds). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form; proxy administration and multisig ownership are outside the supplied file (default-branch snapshot stader-labs/maticX@b27e328).
  • Governance of ValidatorShare and StakeManager contracts has several privileges that can negatively impact MaticX (Bail Security, MaticX, 2024-09). High (estimated: the audit label was "Governance"; the privileges sit in Polygon contracts). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form; the Polygon functions the finding cites are not in the supplied file.
  • MaticX conversion is susceptible to inflation attack (after new deployment) (Bail Security, MaticX, 2024-09). High (estimated: the audit label was "applicable only to forking", but a first depositor could take later depositors' funds). Status: acknowledged. Code check: the default-branch snapshot (stader-labs/maticX@b27e328) may still contain the code this finding describes (low confidence; deployment not confirmed)

Medium

Total groups: 71; fixed per the auditor: 44; not marked fixed: 27.

ETHx and SD token (Ethereum):

  • Slashing of a validator can exceed the actual penalty due (Halborn, 2023-04-24). Status: partially fixed. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0x278097f753ab3e4a5e897dbb72cea1592e78e9a5 (Sourcify runtime match; identical to public commit stader-labs/ethx@1939e6c) (deployed code)
  • Protocol will not benefit from slashing mechanism when remaining penalty bigger than minThreshold (Code4rena, 2023-06-02). Status: acknowledged. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0x278097f753ab3e4a5e897dbb72cea1592e78e9a5 (Sourcify runtime match; identical to public commit stader-labs/ethx@1939e6c) (deployed code)
  • MEV bots can win all the auctions when Auction is paused (Code4rena). Status: disputed. Code check: a fix is visible in the source verified for eip155:1:0xa2aa24e91a345b3dd8652f304390203d0e1c4d31 (Sourcify runtime match; identical to public commit stader-labs/ethx@21d516c); the contract declaration shows no pausing base (deployed code)
  • Owner in VaultProxy.sol is address(0) (Code4rena). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form.
  • ValidatorWithdrawalVault.distributeRewards can be called to make operator slashable (Code4rena). Status: disputed. Code check: disputed by the team; the described code is unchanged in the source verified for eip155:1:0x3073cc90ad39e0c30bb0d4c70f981fbd00f3458f (Sourcify runtime match; identical to public commit stader-labs/ethx@21d516c); not scored (deployed code, low confidence)
  • ValidatorWithdrawalVault.settleFunds doesn’t check amount that user has inside NodeELRewardVault to pay for penalty (Code4rena). Status: acknowledged. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0x3073cc90ad39e0c30bb0d4c70f981fbd00f3458f (Sourcify runtime match; identical to public commit stader-labs/ethx@21d516c); the same holds for the node reward vault at eip155:1:0x97c92752dd8a8947ce453d3e35d2cad5857367af (deployed code, low confidence)
  • No bidder has incentive to bid in the Auction except doing last-minute MEV due to fixed endBlock (Code4rena). Status: acknowledged. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0xa2aa24e91a345b3dd8652f304390203d0e1c4d31 (Sourcify runtime match; identical to public commit stader-labs/ethx@21d516c) (deployed code)
  • Chainlink’s latestRoundData may return a stale or incorrect result (Code4rena 2023-06-02; also reported twice by Halborn, 2023-04-24). Status: acknowledged. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0x1c43cff3abbdae07e57e03f0bd3dd813f40b4785 (Sourcify runtime match; identical to public commit stader-labs/ethx@9d4a921) (deployed code)
  • HasEnoughSDCollateral() Check Is Performed Only Once During Onboarding (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form (deployed code).
  • Submit Functions Are Susceptible To Front Running When Trusted Nodes Are Removed (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the code this finding describes is unchanged in the source verified for eip155:1:0x1c43cff3abbdae07e57e03f0bd3dd813f40b4785 (Sourcify runtime match; identical to public commit stader-labs/ethx@9d4a921) (deployed code)
  • Uninitialized WithdrawDelay (Sigma Prime, 2023-06-01). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form; the audited claimWithdraw and withdrawDelay are absent from the deployed code, which uses a different withdrawal flow.
  • Lack of pause functionality within the OFT Adapter (MixBytes, 2024-07-11). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form (source verified for eip155:1:0x39d0ce7d1509a4f22619d0f666978fbe820665e9, Sourcify runtime match; identical to public commit stader-labs/ethx_oft@6a4ffad); the inherited base contract is not among the supplied files, so an inherited pause cannot be ruled out.

MaticX (Polygon), default-branch snapshot stader-labs/maticX@b27e328, deployment not confirmed:

  • The reentrancy guard is missing in MaticX.requestWithdraw (Immunebytes, 2022-04-25). Status: acknowledged. Code check: a fix is visible in the snapshot (nonReentrant on the withdrawal request)
  • Malicious user can DoS withdrawals by dusting the MaticX contract with small amounts of ValidatorShare tokens from different validators (Bail Security, 2024-09). Status: acknowledged. Code check: the snapshot may still contain the code this finding describes (low confidence)
  • Lack of reward compounding before submit allows users to extract value from the protocol (Bail Security, 2024-09). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form.
  • Exotic edge-case will result in bricked submit function (Bail Security, 2024-09). Medium (estimated: the audit label was "applicable only to forking", and the case can lock deposits). Status: acknowledged. Code check: the snapshot may still contain the code this finding describes (low confidence)

BnbX (BNB Chain), default-branch snapshot stader-labs/bnbX@18442f8, deployment not confirmed:

  • Governance Privilege: Users are completely dependent on the operator to withdraw funds (Bail Security, 2024-07). Medium (estimated: the audit label was "Governance", and users cannot withdraw if the operator does not act). Status: acknowledged. Code check: the snapshot may still contain the code this finding describes (low confidence); starting a batch undelegation remains operator-only
  • Lack of ER update allows users to flash-theft tokens (Bail Security, 2024-07). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form; the recommended change is only partly visible.

NearX, Terra and the December 2021 report (code check unavailable unless stated):

  • Potential Centralization Problem (I) and (II) (BlockSec, NearX, 2022-10-13). Medium (estimated: the audit gave no severity, and the admin can withdraw all wNear tokens). Status: acknowledged. Code check: not marked fixed per the auditor (acknowledged); our code check is unavailable: the audited repository is not available.
  • Four Terra findings (Oak Security, 2021-12-06; Terra is inactive). Status: acknowledged. Code check: not marked fixed per the auditor (acknowledged); our code check is unavailable: the audited repository is not available.
  • Possibility of for loop exceeds block (December 2021 report by The Z Institute; weak source). Status: acknowledged. Code check: the checked version of the code does not contain the described code in this form (NearX snapshot stader-labs/nearx@3912ee1).
  • Two gas-efficiency findings (same report): "Unoptimized loop" and "Unoptimized coding block". Status: acknowledged. Code check: not applicable; these are gas/efficiency findings with no security weight, so no security check was run.

Recurring Patterns

  • Two separate slashing issues sit in the ETHx collateral contract. Halborn (partially fixed) describes slashing more SD collateral than the actual penalty . Code4rena (acknowledged) describes the protocol not benefiting from the slashing mechanism when the remaining penalty exceeds minThreshold, so less may be recovered. The described code is unchanged in the verified Ethereum source for both. The audits do not describe either as a direct drain of user funds.
  • Stale or unchecked oracle data appears in Halborn and Code4rena for the ETHx oracle, and the described code is unchanged in the verified source.
  • Operator and admin privilege recurs outside ETHx: Bail Security on MaticX and BnbX, and BlockSec on NearX. These are acknowledged governance findings; whether they matter depends on who holds the roles, which this report does not assess.
  • Auction design (fixed end block, last-minute bidding) is acknowledged in the Code4rena contest and unchanged in the verified source.
  • Findings on MaticX and BnbX were checked only against default-branch snapshots, not confirmed deployments.

Code Quality Assessment

  • ETHx has the deepest review: Halborn, Sigma Prime, Code4rena and MixBytes between 2023 and 2024. MaticX and BnbX have 2022 reviews plus 2024 Bail Security reviews; NearX and Terra reviews are older, and those chains are inactive or small.
  • Fix rate is high: 79 of 110 critical, high and medium groups are marked fixed, and no critical group is open. The 31 not marked fixed are mostly acknowledged.
  • Traceability is the main gap: no audited revision could be matched to a deployed component, and six fixed ETHx findings sit in files that differ from the reviewed fix revision.
  • Source is verified for the 66 EVM deployments in the verification inventory: 64 Sourcify runtime matches and 2 explorer-listed sources (an explorer claim).
  • Control read: the ETHx permissionless node registry reported paused at Ethereum block 26384083 . Who holds the admin roles is outside this report.
  • Known compiler bugs: the compiler-bug list covers deployed contracts built with 13 Solidity versions from 0.6.6 to 0.8.25, most commonly 0.8.16 (35 contracts) and 0.8.22 (14). For 0.8.16 it lists the medium/high "StorageWriteRemovalBeforeConditionalTermination", which applies only with the Yul optimizer and assembly return or stop , and the medium "UnsoundSpillInMutualRecursion" (also listed for 0.8.20 and 0.8.22), which applies only with viaIR (https://blog.soliditylang.org/2026/07/09/unsound-spill-in-mutual-recursion-bug/). Other entries rated low or low/medium also apply. Whether these conditions hold for each contract was not checked, and this is not a finding.

Score: 4.5/10

Rubric security-score-v2.6; data completeness 0.65

Code security factors, points out of 10 (points)
Audit coverage
5.50 points
Audit depth
3.50 points
Unresolved findings
4.58 points
Assurance
3.50 points
Audit findings not marked fixed (findings)
High · acknowledged
4 findings
Medium · partially fixed
1 findings
Medium · acknowledged
24 findings
Medium · disputed
2 findings
SeverityGroupsFix visible (deployed code)Fix visible (repository only)Described code still there (deployed code)Described code still there (repository only)Not checked or inconclusive
Critical000000
High4 (1 not scored)0001 (1 low confidence or archived)3 (1 not scored)
Medium27 (4 not scored)1 (1 not scored)17 (1 not scored) (2 low confidence or archived)3 (1 not scored) (3 low confidence or archived)15 (1 not scored)

Not scored: disputed groups and groups excluded for a later audit fix or privileged role.

Low confidence or archived: the check was not confident enough, or the repository is archived; the described code may still be there.

Deployed code: the source verified for the live contract. Repository only: the protocol's public repository; deployment not confirmed. “Described code still there” is a limited code signal, not a confirmed vulnerability: no exploit was tested; other contracts, configuration or operations may mitigate it. Findings the auditor marked fixed are trusted, not re-checked. Ask the team whether these are fixed or mitigated. Raw penalties: deployed code 2.16, repository only 0.92, audit status alone 2.34; Unresolved findings loses min(2.16 + 0.92 + 2.34, 10) of 10 points.

FactorLevelEvidenceWeightPoints
Audit coverage (computed)L2supported0.355.50
Audit depth—not disclosed0.203.50
Unresolved findings (computed)L2supported0.304.58
Assurance—not disclosed0.153.50

66 citations · click markers to view evidence

💰

Economic Security

4.5/10

Economic Security Analysis: Stader

Stader's economic security rests on ETHx, the Ethereum liquid staking token, which anchors about 85% of the captured TVL and is backed by real staked ETH. A three-node committee pushes its exchange rate, a Safe Mode can suspend withdrawals with no stated bound, and the one named insurance fund contract has no stated size and no stated user-loss rule. The SD token adds a reflexive layer, because SD is secondary operator collateral and the SD Utility Pool mints claims against SD that are lent back to operators. Mainnet ETHx depth and holder concentration were not measured, so exit liquidity is the least certain part and confidence is moderate.

Stader's DefiLlama record covers ETHx (Ethereum), HBARX (Hedera) and the SD token with its SD Utility Pool, a lending pool where SD is lent to node operators. Fantom, Terra2, Binance and Near hold no current TVL; Stader was deployed there earlier and those deployments are not assessed. NearX, BNBx and MaticX appear only through a redemption-change notice and past incident records.

TVL & Liquidity Profile

The latest captured TVL is $156.73M, split between Ethereum (about 85%) and Hedera (about 15%), and it fell sharply over the last 90 days.

  • DeFiLlama snapshot, collected 2026-11-13T22:39:03Z: . Ethereum holds and Hedera holds .
  • Price effect is separate from outflows. Over 30 days from 2026-10-14, TVL changed by USD -52,989,587, of which USD -41,823,740 is price effect . The remainder, about USD 11.2M at current prices, is a drop in staked quantity.
  • Over 90 days from 2026-08-15, TVL changed by USD -35,260,028 with a price effect of USD +11,085,837, so the quantity decline is about USD 46.3M at current prices .
  • The monthly history ends at USD 162.9M on 2026-11-13, which differs from the snapshot because the two were collected at different times .
TVL by chain, DefiLlama snapshot 2026-11-13 (USD M)
Ethereum
133.0 USD M
Hedera
23.7 USD M

Exit paths.

  • ETHx can be sold on decentralized exchanges or redeemed through the UserWithdrawalManager contract. Holders send ETHx to it, a finalization step declares ETH claimable, and the holder then claims. (Llama Risk report for Prisma, November 2023.)
  • The same report says requests are first-come, first-served with a mandatory delay against sandwich attacks . No source states a maximum delay in days.
  • The official FAQ gives only a typical time: The payout uses the lower of the exchange rate at request and at finalization . Tokens stop earning rewards once the request starts
  • On Ethereum at block 26384083 (2026-11-10), the contract is not paused, and 8,773 is the next request ID against 8,767 as the next to finalize. About 2,710.6 ETH is requested for withdrawal . This indicates a working queue of a few open requests; processing time is not measured.
  • An unverified publisher states that HBARX unstaking takes about 24 hours, set by Hedera's native unstaking time; this is unconfirmed by a stronger source (source reliability: unverified publisher).

Measured market depth. Mainnet ETHx, the main asset, has no depth, holder-concentration or asset-check row in the measured data. The rows that exist cover only minor deployments.

  • ETHx on Arbitrum shows USD 846 of depth through Camelot v3 . Its supply is USD 16,917, and this measured depth is below USD 2M; the mainnet deployment is unmeasured.
  • ETHx on Optimism has no swap route in the depth probe, and SD on Ethereum shows depth 0 . SD depth on BNB Chain is 0 and on Polygon has no route.
  • Only ETHx on Optimism has a measured holder split, taken at block 159467639: the largest holder has 33.61% and the top ten hold 97.73% . Its supply is about USD 423, so the figures show a negligible deployment, not mainnet behaviour.
  • Dated evidence for mainnet: the Llama Risk report (November 2023) puts the combined Curve ETHx pools above $25M . It also says that
  • An open data-collection search read the Curve ETHx/ETH pool at about $189.7k and HBARX at about $5.8M from DexScreener. This is an unverified aggregator read of one pool, not a 2% price-impact measurement, and it sets no level. Mainnet ETHx depth at 2% price impact stays unmeasured.

Concentration risks. The Optimism holder split is the only measured one. The largest concentration fact on mainnet is structural: about 85% of TVL sits in one token.

Oracle Risk

ETHx uses a proprietary, committee-pushed exchange rate protected by on-chain change and frequency limits, with no market price feed of its own; this is adequate for a liquid staking token but depends on three keys.

  • Which oracles. A whitelisted Oracle Committee relays the ETH held by the system to the oracle contract, which sets the ETHx rate . The same oracle carries the SD price used for operator collateral checks .
  • On-chain state. At Ethereum block 26384083 (2026-11-10) the oracle's latest report, for block 26380800, shows 68,711 ETH against 62,387 ETHx, a rate of about 1.1014. The contract is not in safe mode or inspection mode . The change limit is 15 against a basis-point scale of 10,000 . If the limit counts in basis points, each update may move the rate by at most 0.15%; the unit is not confirmed.
  • Committee size. The committee holds three trusted nodes, equal to the contract minimum of three. A minimum was added after an audit found that a single malicious vote could pass with few nodes . That finding is marked fixed.
  • Inspection mode and operator changes. A rate update that jumps outside the allowed range triggers a manual review step . The documentation also promises a brake on committee turnover . The on-chain read shows a cooling period of zero for trusted-node changes , so the documented weekly limit is not visible in that setting. Assessment: the limit may live elsewhere, but it is not confirmed.
  • SD price for collateral. Operators post SD as secondary collateral, so the SD price matters . The committee reports it as a 24-hour average . An audit of the SD Utility Pool found no staleness check on that price . The audit status is closed after an acknowledgement, with no code change recorded in the report. Assessment: a halted committee would leave an old SD price in use.
  • No proof-of-reserve check. The rate does not use Chainlink Proof of Reserve data . Two audits flagged missing staleness checks in that unused path; the team said the path awaits testing . The Llama Risk report (2023) says Stader planned to add such feeds .
  • USD price for other protocols. ETHx has no Chainlink market feed. A Stader oracle contract multiplies the Chainlink ETH/USD price by the internal rate . The rate is a redemption value, not a market price, so a market discount does not show in it.
  • Manipulation vector. A rate set by three keys moves only within the change limit per update. The 2023 report says rate updates are asynchronous and sometimes take over 24 hours, which makes yield and rate lumpy
  • Cross-chain rate. The ETHxRateProvider contract feeds other chains through LayerZero. It stores a rate of 1.0335 with a last-update timestamp of 1719597191 (2024-06-28), while its live read returns 1.1014 . The owner address has no deployed code, so it is a single key; it also sits among the owners of a Stader multisig. A consumer reading the stored value would see a rate about 6% below the live one; which value is served on other chains is not confirmed.
  • Historical incidents. The workspace holds no oracle-caused loss. The DefiLlama hacks and DeFiHackLabs files have no records.

Flash Loan & MEV Exposure

ETHx is not exposed to a flash-loan price attack through its own redemption rate, and the documented MEV risks fall on node-operator behaviour, which is bonded.

  • Flash loans. Assessment: because the rate comes from the committee and not from a pool, a flash loan cannot move it directly. The Immunefi bug bounty keeps these attacks in scope and lists a maximum bounty of USD 1,000,000 . Audits of discontinued products include a flash-theft finding for BNBx, where deposits could capture accrued income before a rate update; Stader argued the 7-day delay made it unprofitable . This is BNBx, not ETHx.
  • Sandwich attacks. The withdrawal queue carries a mandatory delay against sandwiching, as quoted above.
  • MEV theft by operators. Operators could withhold MEV rewards. Stader's simulation says the median MEV per block is about 0.046 ETH and that 99% of blocks pay below 1.29 ETH, which the 4 ETH bond covers with room for penalties . This is a simulation published by Stader.
  • Front-running findings in ETHx audits.
    • A Code4rena audit (June 2023, verified host not confirmed) found that bots could front-run a pause on the SD auction and buy SD at a discount . The sponsor disputed it and removed the pause feature.
    • The same audit found that distributeRewards is permissionless, so an attacker can front-run a withdrawal to shrink the operator share and trigger slashing . The sponsor disputed it as intended behaviour.
  • Mitigations. Stader checks withdrawal credentials before lending 28 ETH to an operator, and imposes a 3 ETH penalty if front-running is detected .

Tokenomics Risk: Yield, Leverage & Reflexivity

ETHx yield is named revenue, Ethereum staking rewards net of a 10% fee, with no leverage in the user position. However, the SD Utility Pool mints user claims against Stader's own token, and the SD emission share of paid yield is not cited.

  • ETHx revenue source. Users earn consensus and execution-layer rewards. The litepaper states that The on-chain fee reads for the live contracts are not part of this evidence.
  • Cited amounts. DefiLlama records fees of USD 352,300 over 30 days and USD 7,122,874 over 365 days, and revenue of USD 20,999 and USD 383,878 . These figures cover the whole protocol, not ETHx alone.
  • HBARX. An unverified publisher's Hedera pages state a 12% protocol fee on staking rewards, which is unconfirmed by a stronger source (source reliability: unverified publisher).
  • Own-token emissions. SD is Stader's own token. The 2023 report says SD rewards liquidity providers and ETHx node operators for locking SD . The SDIncentiveController contract shows an emission of 62010000000000000 raw units per block and a reward end block of 25295524 . That end block lies before the read at block 26384083, so this controller indicates no emissions at the read. Other SD programs were not found. The workspace gives no figure for SD as a share of paid yield.
  • Own token in collateral. Permissionless node operators lock 4 ETH plus 0.4 ETH worth of SD . SD is about 9% of the 4.4 ETH operator collateral and a smaller part of each 32 ETH validator. User claims on ETHx redeem for ETH, not SD.
  • Reflexive path. The audit reports that, in the code it reviewed, operators could hold less SD than required if the SD price falls, because the sufficiency check runs once at onboarding. Whether the current deployment behaves the same is not established . The team accepted this risk. SD collateral is the smaller part of the operator buffer, so the ETH bond carries slashing losses.
  • Leverage. ETHx and HBARX holders take no leverage inside the protocol; the workspace shows no looping product or maximum-LTV setting for users. Any leverage comes from third-party protocols that accept ETHx and is outside this record.
  • SD Utility Pool. SD holders delegate SD to a pool and receive pool shares, and node operators borrow that SD and pay fees on it . The team calls the borrowed balance collateral . The audit describes the delegator step as , and the fix was an initial delegate of 1 SD at initialization. On chain at block 26384083, 7,376,787 SD is utilized against 7,019,296 pool shares outstanding . Utilization is about 90.4% . At the SD price in the depth data, the utilized balance is worth under USD 1M, so the pool is small next to ETHx. Assessment: this design is reflexive. User claims are minted against Stader's own token, which is lent back as operator collateral, so SD price stress hits lenders and the operator buffer together.

Solvency, Liquidation & Bad Debt

ETHx is backed by staked ETH, but the committee reports the backing figure and exits can be suspended with no stated bound. Loss handling relies on the operator bond; an insurance fund contract exists, but its size and any user-loss rule are not stated.

  • Backing. The oracle reports 68,711 ETH against 62,387 ETHx (block 26384083, 2026-11-10), a ratio of about 1.1014, above 1:1 . The ETH total is a committee report, not an independent on-chain read of validator balances. The asset check for mainnet ETHx is missing, so backing against liabilities is not independently measured.
  • Loss handling. A validator's 4 ETH collateral compensates slashing and other network penalties, and operators receive only the remainder after stakers . Stader's own simulation says for the cases it tested. No slashing of ETHx validators had occurred by November 2023
  • Exit suspension. A safe mode disables withdrawals until network conditions stabilise, and the DAO re-enables them . The triggers are a share of validators slashed or over 50% of ETHx validators facing downtime. No time limit is stated. At the research date the contract shows safe mode off and withdrawals unpaused. Assessment: users should treat the redemption path as available with a suspension risk, because the pause powers exist and carry no stated bound.
  • Pause control. The permissionless node registry is paused at the read block , which blocks new permissionless operator registrations but does not touch user exits.
  • Debt positions. ETHx has no borrowers. The SD Utility Pool is the only debt position, and operators are the borrowers. The on-chain liquidation threshold is 70 , and lender exits wait a mandatory delay . That is 50,400 blocks, or 7 days at 12-second blocks. Liquidation appears to be a public call by a liquidator who pays the operator's interest, as the audit scenarios show . No source states a liquidator bonus or auction rule.
  • Utility Pool defects found and fixed. An audit found that liquidation did not clear the debt in some cases . It also found that rounding could create bad debt . Both are marked fixed in commit 21ba418, though the report's host is not verified. A further finding says a position could be wrongly liquidatable .
  • Utility Pool loss order and open risk. The team states the order in which a slashing loss is absorbed . Delegated SD is used last. The audit also warns that the 24-hour SD price could delay liquidations . That finding was closed with an acknowledgement. No on-chain deficit record or risk dashboard was found, so a bad-debt record for the current version is missing.
  • Loss cover. The ETHx repository lists an insurance fund contract at 0xbe3781CE437Cc3fC8c8167913B4d462347D11F20 . The 2023 Llama Risk report says the front-running penalty flows into it . It also says the fund pays 1 ETH to cover a front-run in the permissioned pool . Withdrawals from the contract need a 3-of-6 Safe with no delay , so payouts are discretionary. The documented flows cover operator front-running only, and no non-weak source states that the fund covers user losses, its balance or a payout rule. A third-party summary mentions an insurance fund for incidents (source reliability: unverified publisher).

Incentive Alignment

Operator incentives are aligned through the bond and SD collateral, while ETHx holders carry no emission-driven mercenary risk that the data shows; the main gap is a loss-cover fund of unknown size.

  • Operators. Permissionless operators put up 4 ETH of their own capital plus SD. Persistent underperformance accumulates penalty until the operator is exited and their funds are recycled .
  • Mercenary capital. TVL fell by about USD 11.2M of quantity over 30 days and about USD 46.3M over 90 days (see above). The cause is not stated in the sources. No evidence ties it to the end of an emission campaign.
  • Death-spiral paths. ETHx is a redeemable ETH claim, so a spiral needs mass slashing or a committee failure. SD is an incentive token that backs about 9% of operator collateral at onboarding, although the Utility Pool lends SD back to operators. Assessment: a death spiral of ETHx is unlikely, but an SD price collapse would stress the Utility Pool, based on the structure above.
  • Loss cover and past incidents. The only loss incident with a reimbursement announcement in the workspace is the NearX exploit of 16 August 2022. An attacker minted NearX without staked NEAR, and the total loss was limited to about 165k NEAR, roughly USD 800k . Stader announced it would make affected users whole by covering the loss . The reimbursement plan sets the new NearX at the pre-exploit rate of about 1.0153 . Completion of payments is not confirmed in the sources, and the plan describes no named fund as the payer.
  • Current product status. A 28 October 2026 notice says redemption terms for NearX, BNBx and MaticX are changing; those products are outside the live TVL.

Economic Attack Vectors

The credible attacks target the committee rate, the operator set and the thin secondary-chain markets, not a lending liquidation cascade.

  • Price manipulation. A market price attack on the ETHx rate is not possible because the rate is not a market price. The documentation says oracle operators report the rate by majority consensus . A compromise of a majority of the committee could report a wrong rate within the change limit per update. The committee read shows three nodes, and the exact quorum is not confirmed.
  • Rate-provider key. The cross-chain provider's owner is a single key. The owner's role over the rate served on other chains is not established. The stored rate carries a 2024 timestamp, and which value consumers on other chains see is unconfirmed.
  • Liquidation cascade. Not applicable to ETHx itself. Third-party lending markets that accept ETHx are outside this record, and their exposure to a thin ETHx market is a downstream risk (assessment: moderate, because mainnet depth is unmeasured).
  • Bad debt. No bad-debt record exists for ETHx. For the SD Utility Pool, audits found rounding and liquidation paths that could leave bad debt, now marked fixed, and no recorded deficit was found. At 90.4% utilization, a drop in the SD price relative to operator obligations, combined with the 24-hour price average, would be the first stress point.
  • Slashing correlation. A correlated slashing event could exceed the 4 ETH bond. The safe mode and the pre-signed exit messages exist for this case .
  • Bridge-related risk. Bridged ETHx exists on Arbitrum and Optimism and is small. The mainnet lock pool holds 8.14 ETHx , against about USD 17k of supply on the two chains in the depth and holder data. The mainnet-to-Optimism route uses a Chainlink CCIP rate limit of 1,000 ETHx refilling at 0.056 ETHx per second . The SD route has no rate limit found , but its lock pool holds only 54.79 SD, worth USD 5.36 . A LayerZero v2 adapter route on chain 100 relies on a 4-of-4 verifier set . Chainlink CCIP appears in the dependency index with an edge share of 0.0 and is unresearched. The measured routes have price and coverage gaps, and Chainlink CCIP is unresearched, so the consequence of a verifier or route failure is not bounded here.
  • Critical dependencies. The failure of any of these could cost user principal:
    • Ethereum validators and the node-operator set (a permissionless pool and a permissioned pool).
    • The three-node Oracle Committee.
    • The Hedera network and its staking for HBARX (about 15% of TVL).
    • LayerZero for the cross-chain rate, and CCIP for bridged ETHx.
    • Chainlink ETH/USD for downstream USD pricing.
  • Chain profile. The Ethereum and Hedera profiles have no L2BEAT stage and no exit window ; neither chain is a rollup. The core ETHx contracts were created in June 2023, so the oldest are about 41 months old . The sources reviewed give no per-integration caps for these dependencies.

Aggregator Scores

No CertiK, DeFiSafety or Exponential files exist in the workspace, so there are no curated scores to compare. DeFiLlama is the only aggregator with usable data here: the TVL snapshot above (collected 2026-11-13T22:39:03Z) and the fee history. The DefiLlama hacks file and the DeFiHackLabs file hold no records, which is consistent with the one documented incident (NearX, 2022) being absent from them. That absence does not show that no other incident occurred.

Data Completeness

Available. TVL and its price-effect split, on-chain oracle and withdrawal-queue state at Ethereum block 26384083 (2026-11-10), the 2023 Llama Risk review, official ETHx documentation, audit digests for the ETHx contracts, the SD Utility Pool audit, bridge route settings and one incident record.

Missing or unconfirmed.

  • Mainnet ETHx depth, holder concentration and asset check. The measured rows cover only Arbitrum and Optimism.
  • The insurance fund balance, any user-loss payout rule and payout rules beyond the two front-running cases.
  • The maximum withdrawal delay in days, and the share of SD emissions in paid yield.
  • A liquidator incentive rule and a bad-debt record for the SD Utility Pool.
  • Whether the NearX reimbursement was completed.
  • Several audits come from hosts whose provenance is not verified, which lowers confidence but not the findings.

Dependencies without a usable report. Chainlink CCIP is listed as unresearched with no report, and Chainlink (ETH/USD), LayerZero and Hedera have no entries in the dependency index. The bridge route file is complete for the measured routes, but token prices for WETH and mainnet ETHx are missing for some holders. WETH backing is unresolved.

Confidence. About 0.55: the mechanism, oracle and control facts are supported by official sources and an on-chain read, while market depth, loss-cover size and SD pool bad-debt facts are gaps.

Score: 4.5/10

Rubric security-score-v2.6; data completeness 1.00

Provisional — data incomplete: Liquidity concentration: market depth of the protocol's tokens on every chain they are on. This mark does not change the score.

Economic security factors, points out of 10 (points)
Solvency redemption
3.50 points
Oracle design
5.50 points
Dependency risk
5.50 points
Liquidity concentration
5.50 points
Yield design
3.50 points
FactorLevelEvidenceWeightPoints
Solvency redemptionL1supported0.353.50
Oracle designL2supported0.255.50
Dependency riskL2supported0.155.50
Liquidity concentrationL2supported (provisional)0.105.50
Yield designL1supported0.153.50

90 citations · click markers to view evidence

👥

Team & Transparency

7.1/10

Social & Team Transparency: Stader

Stader is run by an identifiable organisation, not an anonymous team. The sources name a legal entity in Panama and one person, Amitej Gajjala, whom the official docs list only as a Stader DAO member; his co-founder and CEO title appears only on a page from an unverified publisher and is unconfirmed. They also list the nine members of the community multisig, a shared wallet that guards the ETHx contracts. Its one recorded exploit (NearX on NEAR, August 2022, about USD 800k) was disclosed with a post-mortem and a reimbursement plan, but completed repayment is not shown. Public development has slowed: the latest commit is dated 2026-05-18. Evidence is good for team, contracts and audits, and thin for community size and communication cadence.

Team Identity

Team identity is partly public: a named legal entity and one named person (Amitej Gajjala, whose leadership role is unconfirmed), but no full team page in the collected sources.

  • Named person: one page lists "Amitej Gajjala" as "Co-founder and CEO at Stader" (source reliability: unverified publisher) . An older docs page (contracts.md) describes him only as "Member at Stader DAO" . His CEO, co-founder or other leadership role is therefore unconfirmed by an official source. No other founder, CTO or core developer is named in the sources.
  • Community multisig signers: the multisig is described as a 6-of-9 "that includes notable members of the Ethereum community" . Named signers include co-founders of Stakely and Stakin and an Amber Group investment partner . The newer page lists different names than the older one (for example Thorsten Behrens and Karan Sirdesai replace "DefiDad" and Pratik Agarwal), so the signer set has changed over time. The older contracts.md page (official docs) lists the Manager multisig as 3-of-5 and the newer page lists it as 3-of-6 . On-chain on Ethereum at block 26384083 the Manager Safe at 0xAAfb31780e4b9c95Bc920e388f4925A874cd07AF has threshold 3 with 6 owners, which matches the newer page . The Community multisig at the same block shows threshold 6 with 9 owners, matching the 6-of-9 description . The Manager Safe's owner set changed recently: an owner was added on 2026-09-02 . The documented signer names are not matched to on-chain owner addresses here, so the current named signers could not be confirmed. The admin-multi-sig page comes from a publisher the pipeline could not classify (source reliability: unverified publisher), so these signer names are lower confidence than the official docs. Governance scoring covers who can act through these multisigs.
  • Legal entity: a third-party collateral-risk assessment (a collateral-risk assessment with late-2023 data) says "Stader Labs" was founded in April 2021, is registered under the name "STADER LABS" with headquarters in Singapore, and that "Stakeinfra Technologies Inc." operates it, incorporated in Panama . It also says the app's terms of service name Stakeinfra Technologies Inc. as "the Company" . The terms of service themselves are not in the collected sources, so the jurisdiction rests on that assessment.
  • Operating history of the team: the official about page says the protocol "Started out on Terra Blockchain" in November 2021 and reached peak TVL of "$1Bn+ with 40k+ staking wallets" . Biographies and prior projects of individual team members are not in the sources.

Track Record

Stader has operated since late 2021 with one documented exploit, on a NEAR deployment that is no longer active; no hack appears in the incident datasets. Incidents after the supplied material could not be searched, because both web searches failed on a provider quota error.

  • Operating age: the family was first listed on DefiLlama on 2021-12-18, per the code-owned family record. The current Ethereum ETHx contracts were created from June 2023 (for example VaultFactory and PermissionedPool on 2023-06-05) .
  • Current-version core contracts (ETHx on Ethereum): the official ethx repository lists the stake pools manager, user withdrawal manager and the permissionless and permissioned pools .
  • Chains: DefiLlama marks Fantom, Terra2, Binance and Near as inactive chains . Stader was deployed there earlier and those deployments are not assessed. The docs mark Polygon and BNB Chain as discontinued .
  • Incident: NearX exploit, 2022-08-16 (this deployment, NEAR chain). Stader's own post-mortem says an exploit was found on its NEAR liquid staking solution and that the loss was limited to 165k NEAR (about USD 800k) . A bug in Stader's own NEP-141 token implementation let the attacker build a large NearX position without staking and drain DEX pools; Stader paused the contract . A security-research write-up (source reliability: unverified publisher) gives 20 million NearX minted through a loop . Stader stated it would revert NearX to a pre-attack state and make affected users whole . A follow-up plan set out snapshot-based reimbursement and a contract migration at an exchange rate of about 1.0153 . The available sources do not show that reimbursement was completed. TVL on the exact incident day is unknown; the monthly series shows about USD 66.9M on 2022-08-01 , which would put USD 800k near 1.2% of that month-start figure, but this is not a day-of-incident measure. Because incident-day TVL and completed reimbursement are unknown, major is the conservative rule-based default; it does not assert a proven threshold or uncovered user loss. The staking contract itself was reported unaffected. No hack appears in the DefiLlama hacks snapshot or the DeFiHackLabs records for this protocol (both aggregator files list zero records for this protocol).
  • Depeg history of own tokens: the measured ETHx tokens on Arbitrum and Optimism show no depeg runs, but the check was not applicable because no exchange rate was available . A depeg history for the main Ethereum ETHx token was not measured. No par (1:1) claim is cited, so none is asserted.
  • Stress events: the monthly TVL series falls from about USD 718.6M on 2022-05-01 to about USD 51.7M on 2022-06-01 . The sources do not state the cause, so this report does not link it to a market-wide event or to user loss.
  • Roadmap delivery: the about page lists a roadmap of Ethereum expansion, rewards APIs and vaults . ETHx on Ethereum shipped (June 2023 contracts). Delivery of rewards APIs and vaults is not shown in the collected sources.

Social Presence

Public development is sparse in the last year, and no follower or community-size figures are in the sources.

  • GitHub: measured commits fall in 3 of the last 12 months (2025-12, 2026-03, 2026-05); the last commit is 2026-05-18 . Most commits are in stader-node (11 in December 2025). The ethx, maticX, bnbX and nearx repositories are nearly silent. The contributor count is not available.
  • Twitter/X, Discord, Telegram, Reddit: the official site links to all four . Follower counts, member counts and discussion quality are not in the sources: insufficient data.
  • Governance forum: the governance launch post reports "8k+ views, 80+ member sign-ups and 2 proposals" in the forum's first month . Snapshot voting uses a space named staderdao.eth with a 50,000 SD proposal threshold and a quorum of 1 million SD votes .
  • Blog: the blog pages carry "updated on" dates from a mid-2026 crawl, which reflect the crawl and not original posting. Posting cadence is therefore not established.

Transparency Indicators

Contract addresses, audit reports and a bug bounty are public; treasury and revenue reporting by the team is not shown.

  • Address list: the official docs publish an ETHx mainnet contract list, including an insurance fund and the SD token .
  • Audits: the docs name Sigma Prime, Halborn and Code4rena for ETHx . Reports from further firms (PeckShield, Immunebytes, MixBytes, Oak Security, Bail Security and BlockSec) are in the collected audit set. Audit coverage by component is assessed in the code-security report.
  • Bug bounty: the docs state a USD 1M Immunefi bug bounty for ETHx . A captured Immunefi program page for "Stader for ETH" lists a maximum bounty of USD 1,000,000, "Live Since 08 July 2023" and "Last Updated 01 January 2025" . It sets a critical reward of 10% of funds directly affected, with a USD 100,000 minimum . The capture is not a live check at the research date, so present-day status is not confirmed. The docs also say on-chain monitoring via Forta was being implemented .
  • Financials: DefiLlama shows 365-day protocol fees of about USD 7.12M and revenue of about USD 0.38M . The team's own treasury, expense or reserve reports are not in the sources.
  • Open source: the core repositories (ethx, ethx_oft, maticX, bnbX, stader-node) are public under the Stader Labs GitHub organisation. Some audited NEAR repositories were reported unavailable (404), so open-source status is mostly open, with gaps for the discontinued NEAR code.

Red Flags

  • The NearX exploit of 2022-08-16 (about USD 800k) arose from the team's own token implementation. Completed reimbursement is not shown in the sources.
  • Only one person is publicly named, and his leadership role is unconfirmed. Biographies, prior projects and the full team are not in the sources.
  • The legal-entity jurisdiction rests on a third-party assessment, not on the terms of service.
  • The two docs pages list different signer names and different Manager multisig thresholds (3-of-5 versus 3-of-6). The measured on-chain threshold is 3 of 6 at Ethereum block 26384083, but the named signers are not matched to on-chain owners.
  • Development has slowed: 3 of the last 12 months have commits.
  • Several chains (Polygon, BNB Chain, NEAR, Terra) were discontinued or are inactive.

Positive Signals

  • A named legal entity with jurisdiction, one named person (Stader DAO member per official docs; his CEO title is unconfirmed) and a signer set that the docs describe with known industry figures.
  • First-party post-mortem for the one incident, with a stated make-whole plan.
  • Operating since late 2021 with no hack recorded in DefiLlama or DeFiHackLabs for this protocol.
  • Published contract addresses, multiple audit reports and a USD 1M bounty stated in the docs.
  • A community governance path (forum plus Snapshot) was launched in 2022.

Data Completeness

  • Available: official docs, GitHub activity counts, incident write-ups, audit set, contract creation dates, DefiLlama fees and TVL history.
  • Missing: any article or press coverage (no articles in the collected set), Twitter/Discord/Telegram metrics, individual team biographies, confirmation that NearX reimbursement completed, depeg measurement for the main Ethereum ETHx token, and present-day bug bounty status (only a captured Immunefi listing is available). Two web searches for incidents after 2022 both failed on a provider quota error, so none were checked beyond the supplied files.
  • This analysis is based on collected files, not live social media scraping.

Score: 7.1/10

Rubric security-score-v2.6; data completeness 1.00

Team transparency factors, points out of 10 (points)
Team identity
7.50 points
Disclosure
7.50 points
Development activity
5.50 points
FactorLevelEvidenceWeightPoints
Team identityL3supported0.457.50
DisclosureL3supported0.357.50
Development activityL2supported0.205.50

Track record: 6.7/10

Rubric security-score-v2.6; data completeness 0.80

Track record factors, points out of 10 (points)
Operating history
9.50 points
Incident record
6.25 points
Resilience
3.50 points
FactorLevelEvidenceWeightPoints
Operating history (computed)L4supported0.309.50
Incident record (computed)L2supported0.506.25
Resilience—not disclosed0.203.50

42 citations · click markers to view evidence

🗳

Governance

4.9/10

Governance Analysis: Stader

Stader's ETHx core contracts on Ethereum upgrade through a 7-day timelock controlled by a 6-of-9 community multisig, which is a reasonably strong design. Several other powers sit outside that timelock: a 3-of-6 Manager multisig pauses contracts and changes parameters instantly, and it can upgrade the ETHxPriceOracle proxy with no delay. Token voting happens off-chain on Snapshot, the team executes the results, and a handful of wallets cast most votes. Overall this is a multisig-run protocol with limited on-chain community control; the on-chain control data was read on 2026-11-10 and covers Ethereum fully but is marked partial. The analysis rests mainly on ETHx; the available sources give no admin, upgrade or delay data for HBARx on Hedera, so that product's controls stay unassessed.

Governance Structure

Stader is a hybrid: multisigs and a timelock hold the on-chain powers, and an off-chain SD-token vote advises them. The scope here is the Stader family in the workspace. ETHx liquid staking on Ethereum is the largest live product. HBARx liquid staking on Hedera is a second live product: the docs describe it , and DefiLlama attributes about USD 23.7 million of TVL to Hedera against about USD 133.0 million on Ethereum . The workspace holds no on-chain control read, multisig description or timelock data for HBARx, so its admin powers and delays are unresolved and are not covered by the findings below. Earlier MaticX and BNBx deployments are marked discontinued or sunset in the docs, and they are treated as context only here.

  • Type: hybrid. The docs describe three entities that control ETHx: a Community multisig, a Manager multisig and a team-controlled Operator (source reliability: unverified publisher) .
  • Voting mechanism: token-weighted SD voting on Snapshot (space staderdao.eth). A proposal needs a holder of at least 50,000 SD, and the quorum is 1 million SD votes .
  • Execution: votes are not binding on-chain. The docs say "currently proposals are not executable code, they are implemented off chain & executed by the team on a specific chain" . A collateral risk assessment likewise describes governance as "currently conducted off-chain and managed by the Stader multisig" .
  • Council override: a Stader Governance Council with core-team members can veto adverse proposals and pass key proposals that miss quorum, "for the initial few months" . The sources give no sunset date for this power.
  • Timelock: a TimelockController at 0x1112…68b9 has a 7-day minimum delay on Ethereum, read at block 26384083 (2026-11-10). Its delay was set to 604800 seconds on 2023-07-08 . Its only proposer and canceller is the Community multisig, and anyone can execute . No operations were pending, and 45 had been executed; the last executed one was on 2026-06-05.
  • Emergency path: per the Admin Multi-Sig page (source reliability: unverified publisher), the Manager multisig can pause and the Community multisig can unpause and upgrade through the timelock. The Manager multisig's 3-of-6 threshold is confirmed on-chain (see Admin Key Analysis) .
  • Vote concentration (venue snapshot:staderdao.eth): the last five proposals drew only 4 to 9 voters each. The top 10 voters hold 87.28% of the votes cast, and the top voter holds 19.69% . The measured quorum share of supply is not available (null). The venue's quorum is 1,000,000 SD .
  • Recent proposals (the data shows the leading choice, not a pass/fail result):
    • Sunset BNBx, 2026-01-23, leading choice YES with 100% of 4 votes: link. Separately, the docs say Stader DAO is sunsetting BNBx (source reliability: unverified publisher) .
    • What should be done with bought-back SD, 2025-09-03: link.
    • SD Utility Pool utilisation rate change to 4%, 2025-09-03: link.
    • Sunset ETHx native staking on Arbitrum, 2025-03-25: link.
    • SD Utility Pool utilisation rate change, 2025-03-01: link.
    • The sources show no proposal that harmed users.

Admin Key Analysis

Stader splits control across a timelocked community multisig, a Manager multisig that acts instantly, and a team Operator key. The measured Safe thresholds range from 3 to 6 signatures, and the SD bridge pools on Ethereum have an owner whose type and threshold were not read.

Who holds what (Ethereum, block 26384083, 2026-11-10):

  • Community multisig 0x45B9…9fF6 is a Safe with threshold 6 of 9 owners, no modules and no guard . The Admin Multi-Sig page (source reliability: unverified publisher) describes it as a 6-of-9 that includes "notable members of the Ethereum community" . The same page lists its members by name, including Stader's CEO . The on-chain owner list changed in June 2026 (four owner swaps between 2026-06-12 and 2026-06-16). The page gives names only, so the current on-chain owners could not be matched to it. Which current signers are independent of the team, and how many, is therefore unconfirmed.
  • Manager multisig 0xAAfb…07AF is a Safe with threshold 3 of 6 owners and no modules . The Admin Multi-Sig page names no signers. Its owners changed repeatedly, with a latest recorded swap on 2026-09-02. It owns the ProxyAdmin of ETHxPriceOracle (below).
  • Operator 0x8F30…737E is controlled by the Stader team, per the same page (source reliability: unverified publisher) . On-chain data shows it as a single key that co-holds a few instant configuration powers with the Manager multisig, such as updateWithdrawnKeysBatchSize and updateInputKeyCountLimit.

What the keys can do (from the measured privilege table):

  • Upgrade ETHx core proxies: the proxy admin 0x67b1…c86e is owned by the timelock, so upgrades of StaderStakePoolsManager, ETHx, UserWithdrawalManager, StaderConfig, the pools and registries and others take at least 604800 seconds (7 days) . DEFAULT_ADMIN_ROLE on these contracts is also held by the timelock.
  • Mint: ETHx MINTER_ROLE is held by StaderStakePoolsManager, and BURNER_ROLE by UserWithdrawalManager. These are contracts, not keys.
  • Pause: the Manager Safe can pause 11 listed contracts instantly. Unpausing needs the timelock (7 days) on 7 listed contracts.
  • Parameters: 30 configuration functions on StaderConfig and related contracts sit behind the timelock. Another 14 (for example updateMaxDepositAmount, updateMinDepositAmount, updateStaderTreasury) and 8 access-type functions (such as activateNodeOperator) are callable by the Manager Safe with no delay.
  • Fund movement: the Manager Safe can call withdrawFunds on the RatedOracle contract and withdrawFund on one other listed contract with no delay. The parser did not resolve what funds those hold. Withdrawal-related settings such as updateUserWithdrawManager are behind the timelock.
  • Oracle upgrade bypass: ETHxPriceOracle (0x9305…B2F) upgrades through ProxyAdmin 0xd99e…f898, owned by the 3-of-6 Manager Safe, with no delay . This oracle is a cross-chain ETHx price source, not the core minting path; the measured file does not show what depends on it.
  • Other chains: ETHx on Arbitrum, Optimism and Gnosis upgrades through a ProxyAdmin owned by a 3-of-5, 3-of-6 or 3-of-5 Safe respectively, with no timelock in the proxy-admin path . Arbitrum, Optimism and Gnosis blocks are 522958195, 159354986 and 49206294 (all read 2026-11-10). On Arbitrum, a 3-hour timelock (10800 seconds) with a 6-of-12 proposer Safe owns the BURNER_ROLE holder of ETHx; what else it controls was not traced.
  • SD token bridge pools (Chainlink CCIP, own token): the measured bridge data lists pools for the SD token with these owners. On Ethereum (block 26393313), the pool owner is 0x4483…9449 . The onchain control file holds no read of this address, so its type, threshold and delay are unknown. On Optimism (block 159410482), the owner is 0x1e2f…a45, a TimelockController with a 3-hour minimum delay (control data read at block 159354986) and a 6-of-12 Safe among its three proposers [aggregators/bridge_routes.yaml:"chain: eip155:10 owner: '0x1e2faed390be7945acbf22fc25173c4bdda3ea45' token: '0x32e94bb4a3521ebebe14ae3aa5bc09fcc9e42306'"; aggregators/onchain_control.yaml:"timelock: style: oz min_delay_seconds: 10800"; "modules: [] threshold: 6 owners_count: 12 modules_truncated: false"]. On Arbitrum (block 523347281), the owner is 0x8a89…c7f, the same 3-hour timelock pattern (control data read at block 522958195) with a 6-of-12 Safe among its three proposers . The other two proposers on each timelock have no Safe data, so their type is unknown. The route data reports no owner threshold for these pools and no delegate. These pools bear on SD token transfers between chains, not on the ETHx core upgrade path.
  • Rate provider: ETHxRateProvider is owned by 0xc616…579a, a single address with no Safe data that is also listed as one of the Community multisig's owners.

Assessment: the admin powers fit a liquid staking protocol that still needs emergency pauses. The community multisig plus 7-day timelock is a real safeguard for core upgrades: 6 of 9 signatures are needed and 7 days pass before execution. How many of the 9 signers are independent of the team is not confirmed, because the page's member list could not be matched to the current owners after the 2026 signer changes, and no other source names the current signers. The weak points are the 3-signature Manager Safe, which acts with no delay, and the lack of identity data for its signers. A pause-capable fast key is useful in an emergency, but here the same key also holds instant parameter powers and the oracle-proxy upgrade.

Who can upgrade or pause ETHx core contractsOpen full view →
ACTORCommunity multisig6-of-9CONTRACTTimelockController 7daysCONTRACTETHx core proxiesACTORManager multisig3-of-6CONTRACTETHxPriceOracle proxyproposes andcancelsupgrades after 7daysupgrades with nodelay
  • (actor)
  • (contract)
  • (contract)
  • (actor)
  • (contract)
  • control
  • control
  • controlrisk

Upgradability

ETHx contracts are upgradeable proxies. Core upgrades pass a 7-day timelock, but some satellite contracts do not.

  • Pattern: transparent proxies with an EIP-1967 admin. On Ethereum, StaderStakePoolsManager at 0xcf5e…d299 has implementation 0x716d…1ee9 and admin 0x67b1…c86e. The measured upgrade list for it shows three upgrades, the latest on 2023-10-16 .
  • Who upgrades: the timelock owns the proxy admin for the main ETHx contracts, and the Community multisig is its only proposer. Contract docs say the timelock "assumes ownership of all deployed ETHx contracts" with a "minimum proposal delay of 7 days" . The on-chain read confirms the 7-day delay and the timelock's ownership for the listed core proxies at block 26384083. The docs' claim that it owns "all" ETHx contracts is not confirmed: ETHxPriceOracle and the other-chain ETHx proxies are exceptions (see above).
  • Latest recorded upgrades (Ethereum): PermissionlessNodeRegistry 2026-01-14; StaderConfig 2025-04-15; StaderOracle 2024-11-13; ValidatorRegistry 2024-11-25 (5-minute timelock path, discontinued MaticX infrastructure). None of the recorded upgrade rows carries a timelock transaction field, so the file does not tie these to the timelock's history; that alone does not prove a bypass, since the timelock had 45 executed operations.
  • Invariant risk: an upgrade can rewrite token and withdrawal logic. For the timelocked contracts, users get at least 7 days of notice and the Community multisig must act. A Manager-Safe upgrade of the oracle proxy has no notice.
  • Immutable parts: ETHx_OFTAdapter on Ethereum and ETHx_OFT on Arbitrum and Optimism show owner 0x0000…0000 in the measured data, meaning ownership is renounced. The Gnosis ETHx_OFT is owned by the 3-of-5 Safe. This does not show the proxies behind them are immutable.
  • Paused state: at the read blocks, PermissionlessNodeRegistry (0x4f4b…fcf1) on Ethereum was paused, and the BNB Chain StakeManager was paused. Neither pause is explained in the sources. BNB Chain StakeManager upgrades pass a 3-day timelock (259200 seconds) with a 4-of-7 proposer Safe.

Parameter Controls

Most fund-affecting settings sit behind the timelock, but deposit limits, treasury address, rewards thresholds and several node-operator settings can change instantly.

  • Timelocked (7 days): 30 configuration functions (such as updateStaderConfig, updateAuctionContract, updateETHxToken), withdrawal parameters (updateMaxWithdrawAmount, updateMinWithdrawAmount, updateMinBlockDelayToFinalizeWithdrawRequest, updateUserWithdrawManager), updateStaderOracle, and updateAdmin.
  • Instant, 3-of-6 Manager Safe: 14 configuration functions including updateMaxDepositAmount, updateMinDepositAmount, updateStaderTreasury, plus updateRewardsThreshold and node-operator controls (activateNodeOperator, deactivateNodeOperator, updateMaxKeysPerOperator).
  • Instant, Manager Safe or Operator key: updateInputKeyCountLimit, updateVerifiedKeysBatchSize, updateWithdrawnKeysBatchSize, updateMaxOperatorId.
  • Oracle data submitters: functions such as submitExchangeRateData and submitSDPrice show no resolved holder in the privilege table (the parser did not resolve who calls them), so who feeds the ETHx exchange rate is not confirmed here.
  • Bounds and harm: the sources reviewed here do not state numeric caps on these parameters. A malicious or compromised Manager Safe could pause deposits and withdrawals or change the treasury address. It could not change the withdrawal-delay setting without the timelock. The exit window (time a user needs to withdraw fully) is not quantified in the sources reviewed; the docs only say each withdrawal request faces "a mandatory delay" .

Decentralization Assessment

ETHx core upgrades need 6 of 9 signatures and a 7-day delay, while operations and voting are centralized. Whether this is more decentralized than a team-only multisig cannot be established, because the current signers' independence from the team is unconfirmed; it is less decentralized than binding on-chain token governance.

  • Roadmap: the Admin Multi-Sig page (source reliability: unverified publisher) states a future vision of "transitioning towards complete on-chain governance, enabling SD and ETHx holders to actively propose and implement protocol changes" . The workspace shows no on-chain governance contract and no date.
  • Key-person risk: a single signer cannot upgrade the ETHx core, since that needs 6 of 9 signatures and 7 days. The Stader team can act alone through the Operator key and, if it controls three Manager signers, through the Manager Safe. The identity of Manager signers is not documented.
  • Token distribution: the SD holder-concentration read for Ethereum SD did not produce a usable top-10 figure (status unverified or error in the file), so SD distribution is not confirmed. Voting power at the Snapshot venue is concentrated: 19 distinct voters in the measured window and an 87.28% top-10 share.

Governance Risks

  • Manager Safe single point: three signers can pause the protocol, change deposit limits and treasury address, and upgrade the oracle proxy with no delay.
  • Off-chain votes: nothing in code forces the team to follow a Snapshot result. The Governance Council can veto or pass proposals, and has no sunset date in the docs.
  • Concentrated voting: with 4 to 9 voters per proposal and an 87.28% top-10 share, a few wallets can pass or block proposals. Low participation also makes the 1,000,000 SD quorum a thin barrier. The 4-voter BNBx sunset vote carried 1,253,706.69 SD of votes .
  • Signer churn: both multisigs changed owners in 2026, and the Admin Multi-Sig page's member list is not matched to current owners.
  • Cross-chain proxies: ETHx on Arbitrum, Optimism and Gnosis upgrades through 3-signature Safes with no timelock in the path.
  • SD bridge pools: the Ethereum pool owner is unread, and the Optimism and Arbitrum pool owners act after a 3-hour delay.
  • Audit findings: audits flagged similar admin centralization. An audit finding says "the extensive authority held by the ADMIN account, including potential fund misappropriation from users and operators, combined with the system's extensive upgradability" . The audit scope and fix status for the current deployment were not compared here.
  • Mitigating factors: a 7-day public timelock behind a 6-of-9 community multisig whose current signer independence is unconfirmed, and fast pause response for emergencies.
  • HBARx: admin keys, upgrade path and delays for HBARx on Hedera are unknown from the available sources.

Data Completeness

  • Available: on-chain control for Ethereum (partial status; block 26384083), Arbitrum, Optimism, Gnosis, Polygon and BNB Chain (read 2026-11-10); Snapshot vote concentration for staderdao.eth; Stader docs on the multisigs, Snapshot process and council. SD token bridge-pool owners come from the bridge-route data and the Optimism and Arbitrum owners were also read in the control data. The Admin Multi-Sig page has source reliability unverified publisher, so its multisig descriptions are confirmed only where on-chain reads agree (thresholds and owner counts).
  • Missing: all HBARx control data on Hedera (a web check for it failed), the type, threshold and delay of the Ethereum SD pool owner, four Arbitrum Sepolia contracts were not read (testnet). Manager Safe signer identities are undisclosed. The current identities of the Community multisig owners were not matched to the docs list. The SD token top-10 holder share could not be computed. The ETHx withdrawal exit window is not quantified in the sources read. The reason for the two paused contracts is not stated. Whether the instant Manager Safe powers over the oracle proxy affect minting or redemption was not traced.
  • Audit provenance: many Stader audit copies are hosted on protocol-controlled or unverified hosts; this lowers confidence in audit citations but is not a penalty.

Score: 4.9/10

Rubric security-score-v2.6; data completeness 1.00

Governance factors, points out of 10 (points)
Admin control
5.50 points
Change delay
3.50 points
Decision process
5.50 points
FactorLevelEvidenceWeightPoints
Admin controlL2supported0.455.50
Change delayL1supported0.303.50
Decision processL2supported0.255.50

40 citations · click markers to view evidence

🔗

Cross-Analysis

Cross-Risk Analysis: Stader

Stader's biggest combined risk is not a code bug. One fast 3-of-6 Manager multisig holds the pause switch, the instant upgrade of the ETHx price oracle and the withdrawal key to the insurance fund, while its signers are undisclosed. ETHx on Ethereum (about $133.0M, 85% of TVL) is backed by staked ETH and has a solid 7-day timelock for core upgrades. The cross-domain gaps sit around it: unbounded exit suspension, a three-node oracle committee, a loss-cover fund of unknown size, and HBARX on Hedera (about $23.7M, 15% of TVL) with no visible audit, controls or mechanics. Confidence is moderate: mainnet ETHx market depth was not measured, and no audit could be tied to the deployed code version.

Stader is a liquid staking protocol, not a bridge, so the "Conditions of Safe Use" section does not apply. Its products are ETHx (Ethereum), HBARX (Hedera) and the SD governance token with the SD Utility Pool. Fantom, Terra2, Binance and Near are inactive chains with no current TVL; Stader was deployed there earlier and those deployments are not assessed .

TVL by chain in USD millions, DefiLlama snapshot 13 Nov 2026 (USD m)
Ethereum (ETHx)
133.0 USD m
Hedera (HBARX)
23.7 USD m

Top Risks for a Holder

The three most serious risks are an exit freeze behind the fast Manager key, a committee-set and instantly upgradable ETHx rate, and an unassessed HBARX product.

1. ETHx exits frozen or redirected through the fast Manager multisig (ETHx, about $133.0M)

  • What can go wrong: three of the six Manager signers can pause contracts at once, but unpausing goes through the 7-day timelock. The Manager signers are not identified, and the owner set changed repeatedly, most recently on 2026-09-02. A separate, protocol-defined freeze exists too: safe mode disables withdrawals, and "No time limit is stated."
  • Cross-domain link: the economic report treats the redemption path as available with a suspension risk, and the governance report shows who can trigger that suspension and how long it takes to reverse. Normal unstaking already takes "typically processed within 7–10 days". Unpausing the 7 listed contracts goes through the 7-day timelock. The sources do not show how that delay overlaps with the 7 to 10-day queue, or how long any given exit would be blocked, so the combined duration is unknown. The same key also changes deposit limits and the treasury address.
  • Magnitude: a freeze could block the ETHx redemption path for up to $133.0M; unpausing needs the 7-day timelock, and the safe-mode duration is not stated, so the length of a freeze is unknown. The reports show no path for it to take principal directly: the Manager Safe cannot change the withdrawal-delay setting without the timelock.
  • Likelihood: low. Basis: it needs three compromised or colluding Manager keys, or a mistaken pause; no pause-caused harm is recorded, and the community multisig (6-of-9, 7-day timelock) can unpause. The fact that would change the band is the identity and independence of the Manager signers.

2. ETHx exchange rate and downstream price set by three keys and an instantly upgradable oracle (ETHx; third-party integrators)

  • What can go wrong: a three-node committee reports the ETH backing that sets the ETHx rate. The ETH total is "a committee report, not an independent on-chain read of validator balances". The on-chain cooling period for trusted-node changes reads zero.
  • Combined finding (visible only across reports): the governance report says the ETHxPriceOracle proxy upgrades through the Manager Safe with no delay but does not show what depends on it. The on-chain read shows the cross-chain ETHxRateProvider points at that same oracle address , and the proxy upgrade needs no delay . The economic report says a Stader oracle contract multiplies Chainlink ETH/USD by the internal rate to give other protocols a USD price. The reports do not state that this is the same contract as ETHxPriceOracle. The sources do not show whether the USD oracle is ETHxPriceOracle or which lenders or chains read either contract, so the downstream effect of a 3-signature upgrade with no delay is unestablished. The measured fact is that the oracle the rate provider points at can be upgraded by three Manager signers at once. Tracing the consumers of that address would decide the downstream exposure.
  • Weak spots on the rate itself: the committee rate moves per update by at most 0.15% if the limit counts basis points; the unit is not confirmed. On the 68,711 ETH figure, 0.15% is about 103 ETH per update (our arithmetic, conditional on that unit). The provider's stored rate is about 6% below the live one, and its owner is a single key that is also a Community multisig owner.
  • Code evidence: audit findings on stale or unchecked oracle data appear for the ETHx oracle, and the described code is unchanged in the verified source. The economic report places the staleness checks in a proof-of-reserve path that the on-chain read shows is switched off (isPORFeedBasedERData: false), so the flagged path is not the live rate path. This is a limited code signal, not a confirmed vulnerability. Six Sigma Prime fixes cannot be confirmed as deployed as reviewed.
  • Magnitude: the committee's per-update rate limit would bound a rate error only if the limit counts basis points, and its unit is not confirmed. That limit applies to committee updates; the reports do not show that it constrains a malicious upgrade of the separate price oracle proxy, so a loss from that path is not bounded here. For third-party markets that accept ETHx, the exposure cannot be sized: their size is unknown. The measured mainnet lock pool holds 8.14 ETHx and the priced Arbitrum and Optimism supply is about USD 17k, but Gnosis is unpriced and the route file has price and coverage gaps, so cross-chain exposure is not bounded.
  • Likelihood: low. Basis: needs a committee majority or three Manager keys; inspection mode and a minimum-node count exist; no oracle-caused loss is on record. The missing facts are who may add or remove trusted nodes, the unit of the change limit, and which contract integrators read.

3. HBARX on Hedera is unassessed (HBARX, about $23.7M, 15% of TVL)

  • What can go wrong: four contract IDs are listed, but the reports show no mechanics, controls or audit. Admin keys, upgrade path and delays are unknown. Stader's Hedera page links a Halborn audit, but the report could not be retrieved, so its date, scope and applicability are not known. The Immunefi bounty evidence does not show a Hedera program.
  • Why it only shows up in combination: all four domain reports leave the same hole. Code, governance and economic reviews each cover only ETHx, so a 15% slice of TVL has no review in any domain. An unverified publisher states a 12% reward fee and about 24 hours to unstake; neither is confirmed by a stronger source. (source reliability: unverified publisher) (source reliability: unverified publisher)
  • Magnitude: up to about $23.7M if the contracts are exploitable or the keys misused; loss of funds is the worst case.
  • Likelihood: unknown. Basis: the reports hold no audit, source, admin-key or upgrade-delay data for HBARX, so no band can be supported. A likelihood assessment of contract or key failure needs the Hedera admin keys and upgrade path, the deployed contract source, and the date, scope and findings of the linked Halborn audit.

Combined Attack Vectors

Vector 1: Fast-key compromise (pause, oracle upgrade, fund withdrawal)

  1. An attacker compromises or colludes with three of the six Manager Safe signers; their identities are undisclosed and the owner set changed in 2026.
  2. The attacker upgrades the ETHxPriceOracle proxy at once, which the on-chain read shows needs no delay. The cross-chain ETHxRateProvider points at that oracle address; which integrators read it is not traced in the reports.
  3. The attacker pauses ETHx, the pool manager, the withdrawal manager and the oracle. Unpausing needs the 7-day timelock, proposed by the Community multisig.
  4. The same Safe is the holder of withdrawFund(uint256) on the insurance fund contract, with no delay. The economic report concludes payouts from it are discretionary; the governance report left this function unresolved, and the on-chain file ties both to one Safe.

Domains: governance (fast keys) + economic (oracle use, loss cover) + overview (pause design). Code audits do not close it, because the audits could not be tied to the deployed versions. Likelihood: low. Basis: needs three keys; Safe thresholds and a 6-of-9 unpause path exist; no such event on record. Impact: severe if the upgraded oracle feeds a consumer that holds value, which is not traced; moderate for the core pool, whose main upgrade path stays behind the 7-day timelock.

Vector 2: Correlated slashing with an unbonded permissioned pool and a discretionary loss fund

  1. A shared client bug or correlated operator failure slashes many validators. Stader's own simulation says a 4 ETH bond covers the tail risks it tested; the economic report notes "A correlated slashing event could exceed the 4 ETH bond."
  2. Permissioned operators post no bond, so the overview finds that sources do not show what absorbs their losses.
  3. The slashing logic in the collateral contract carries two audit findings (Halborn, partially fixed; Code4rena, acknowledged), with the described code unchanged in the verified source; neither is described as a direct drain of user funds.
  4. The insurance fund's documented flows cover operator front-running only, with no stated size or user-loss rule, and the Manager Safe controls its withdrawal key.
  5. Safe mode suspends withdrawals while the loss is crystallising, so holders cannot exit during the event.

Domains: economic (loss waterfall) + code (slashing findings) + governance (fund key, pause) + overview (unbonded permissioned pool). Likelihood: low. Basis: no slashing of ETHx validators had occurred by November 2023 per the Llama Risk report; that report is dated and a later record was not found. Impact: severe if it happens (loss of principal for all ETHx holders pro rata), because user loss absorption beyond the 4 ETH bond is undefined.

Vector 3: SD price stress hits the Utility Pool and the operator buffer together

  1. SD falls sharply; it has measured depth of 0 on the probed Ethereum route and a market cap of about $7.35M.
  2. The SD price used for collateral is a 24-hour average reported by the same three-node committee; an audit found no staleness check, so a halted committee leaves an old price in use.
  3. Operators can hold less SD than required, because the sufficiency check runs once at onboarding. The audit reports this for the code it reviewed; the code report found the described code is not present in this form in the deployed code, so current behaviour is not established.
  4. At about 90.4% utilisation, delayed liquidations could leave bad debt in the pool; lenders and the operator buffer both lose value. On chain at block 26384083, 7,376,787 SD is utilised against 7,019,296 pool shares, and lender exits wait a mandatory 50,400 blocks, about 7 days. These are outstanding balances, not measured bad debt. Slashing losses hit operator ETH and self-bonded SD first, and delegated SD last, which limits the chance that lenders absorb a loss.

Domains: economic (reflexivity) + code (Sigma Prime acknowledged findings) + oracle (committee-set price). Likelihood: medium. Basis: a small token with no measured depth can move sharply, and the audit findings were acknowledged rather than fixed; no bad-debt event is on record. Impact: moderate. The utilised balance is worth under USD 1M, so the pool is small next to ETHx, and ETHx holders redeem for ETH, not SD.

Vector 4: Run on ETHx with thin, unmeasured exit liquidity

  1. Holders leave: TVL quantity fell by about USD 11.2M over 30 days and about USD 46.3M over 90 days, and the cause is not stated.
  2. Mainnet ETHx depth at 2% price impact is unmeasured; the only dated mainnet data is a November 2023 figure (Curve pools above $25M) and an unverified aggregator read of one Curve pool at about $189.7k.
  3. The protocol route takes typically 7 to 10 days, pays the lower of the request-time and finalization-time rate, and stops rewards.
  4. If safe mode or a Manager pause hits during the run (Vector 1), the secondary-market discount would have no redemption anchor while the pause lasts; unpausing needs the 7-day timelock, and the safe-mode duration is not stated.

Domains: economic (liquidity, TVL trend) + governance (pause asymmetry) + overview (queue mechanics). Likelihood: medium. Basis: outflows are already visible, exit depth is unmeasured; only a discount, not loss of principal, is the likely form. Impact: moderate (a temporary secondary-market discount or delayed exit; no loss of backing).

Vector 5: A public core upgrade whose notice is shorter than the exit queue

  1. An attacker compromises six of the nine Community multisig signers and schedules a harmful core upgrade. The timelock gives 7 days of notice. The upgrade can rewrite token and withdrawal logic.
  2. Holders who start the protocol redemption on day 0 face a queue that is typically 7 to 10 days, so some requests may not finish before the upgrade executes. Selling ETHx on the mainnet market is the other route, and its depth is unmeasured.
  3. The Community multisig is also the timelock's only proposer and canceller, so the only party able to cancel a malicious proposal is the compromised one. A Manager pause is a separate authority; compromise of one Safe does not show compromise of the other.

Domains: governance (timelock) + economic (queue length, market depth). Likelihood: low. Basis: six signatures plus public notice are substantial barriers, and no such event is on record. Signer independence is the decisive unknown, and the owner set changed in 2026 while the named signers could not be matched to current owners. Impact: catastrophic for the core ETHx pool (up to about $133.0M), because the harmful code would be the core token and withdrawal logic; the 7-day notice reduces surprise but does not guarantee that exits finish.

Vector 6: Satellite proxy-admin compromise meets thin local liquidity

  1. ETHx on Arbitrum, Optimism and Gnosis upgrades through a ProxyAdmin owned by a 3-of-5, 3-of-6 or 3-of-5 Safe, with no timelock in the proxy-admin path.
  2. A harmful proxy change would hit local ETHx holders, who have few ways out: Arbitrum shows USD 846 of depth against USD 16,917 of supply, and Optimism has no swap route in the depth probe.
  3. The LayerZero v2 route to Gnosis needs all four configured verifiers and reports a config-change delay of 0.0 hours, so a verifier outage halts that route and a config change takes effect at once.

Domains: governance (satellite proxy owners) + bridge settings + market liquidity. Likelihood: low. Basis: needs three compromised keys on a satellite Safe; no consumer of the satellite tokens is traced. Impact: moderate for the priced satellite supplies only (about USD 17k on Arbitrum and Optimism); Gnosis is unpriced and route coverage has gaps, so the total satellite impact is not bounded. The reports show no path from a satellite failure to the Ethereum ETHx backing.

Systemic Risk Factors

  • Dependencies (route to the economic dependency factor): the dependency graph lists Stader as backed by WETH with share 1.0, and that token is "unresolved". Chainlink CCIP is a bridge dependency with share 0.0, status unresearched, no published report and no score. The graph's warnings list is empty as of 2026-11-13: no circular backing or downstream holders are flagged. Chainlink ETH/USD, LayerZero and Hedera have no entries in the dependency index.
  • Bridges carrying the protocol's own tokens: the CCIP edge is 0.0 of TVL. The measured mainnet lock pool holds 8.14 ETHx, and priced ETHx supply on Arbitrum and Optimism is about USD 17k; Gnosis is unpriced and route coverage has gaps, so total bridged exposure is not bounded. The mainnet-to-Optimism ETHx route has a set rate limit; the SD routes have none found, but their lock pool holds only 54.79 SD (USD 5.36). The LayerZero v2 mint-burn route on Gnosis reads: with a config-change delay of 0.0 hours. The CCIP routes have null verifier settings, which are unknown, not met. The CCIP SD settings recur on all measured directed routes: shared defaults, receive settings not pinned, and a reported config-change delay of 3.0 hours. The 3-hour figure is not verified owner-control assurance on Ethereum, because the Ethereum SD pool owner 0x4483…9449 was not read. The ETHx CCIP route has a rate limit of 1,000 ETHx refilling at 0.056 ETHx per second, which is large next to the 8.14 ETHx lock balance, so it is not a tight cap on the current balance. Selected CCIP OffRamps report a 6-of-16 quorum, but other OffRamps were unread, so that cannot stand in for the route-wide quorum. No published bridge report exists to compare these settings against. The route file status is completed.
  • Chain-level: 84.85% of TVL lives on Ethereum and 15.14% on Hedera. Neither chain is a rollup, so there is no sequencer risk; neither has an L2BEAT stage.
  • Macro: most of the TVL is ETH-denominated, so USD TVL moves with the ETH price: of the 30-day fall of USD 52,989,587, USD 41,823,740 is price effect. A mainnet ETHx depeg history was not measured.
  • Correlation: an ETH price fall, a safe-mode freeze and a thin SD market can hit together (Vectors 3 and 4). Ethereum and Hedera positions share no contract or key in the reports, so those two failures look independent; the HBARX control set is unread, so this is not confirmed.

Risk Amplifiers

  • Instant pause + 7-day unpause: an error or compromise can extend the time exits are blocked, because unpausing needs the 7-day timelock; the total delay for a given exit is not shown in the sources (Risk 1 and Vector 4).
  • Undisclosed Manager signers + signer churn + instant powers: the identity of the key set is not known while the powers are broad (Vector 1). The Operator role is a single key without deployed code that co-holds some instant configuration powers.
  • Audit traceability gap + deployed code with unchanged flagged patterns: with no audit tied to a deployed version, the fix status of the findings cannot be confirmed (Vectors 1 and 2). The coverage table has 33 deployed components, with 14,747 lines in 90 verified files, all in the version-not-compared bucket. That is an unperformed comparison, not a measured mismatch.
  • Notice shorter than the exit queue: the 7-day timelock notice is shorter than the 7 to 10-day protocol redemption, and Safe Mode has no stated maximum, so public warning does not guarantee an exit (Vector 5).
  • Provenance of some audit copies: several audit files are hosted by the protocol or by unverified hosts, which lowers confidence in their provenance without changing the finding assessments.
  • Three-node committee reporting both the ETHx backing and the SD price: one small key set underlies both the main token and the SD pool (Risk 2 and Vector 3).
  • Low voter turnout: 4 to 9 voters per proposal and an 87.28% top-10 share mean a few wallets decide off-chain votes, which the team then executes.
  • Slow development: commits in 3 of the last 12 months, with the last on 2026-05-18, so acknowledged findings may wait long for a fix.

Risk Mitigators

  • Core upgrade control: ETHx core proxies upgrade only through a 7-day timelock whose only proposer is a 6-of-9 Community multisig. The timelock's delay is confirmed on-chain at Ethereum block 26384083.
  • Audit record across products: Halborn, Sigma Prime, Code4rena and MixBytes reviewed ETHx. The count of 79 fixed of 110 critical, high and medium groups, with no critical group open, spans all products and audits, including inactive ones, so it is not an ETHx-specific figure.
  • Product-specific reviews: Halborn reviewed ETHx between April 24 and July 4, 2023, MixBytes reviewed ETHx.sol and the cross-chain adapter in July 2024, and Sigma Prime's January 2024 review covers the SD Utility Pool only, not HBARX. The SD review was limited to the changes in PR 212.
  • First-loss design for permissionless operators: operators post a 4 ETH bond plus 0.4 ETH worth of SD, and pre-signed exit messages let Stader force-exit a validator before penalties get dangerous. Permissioned operators have no equivalent stated bond.
  • Verified source: 64 Sourcify runtime matches and 2 explorer-listed sources for the 66 EVM deployments.
  • Backing and exit design: ETHx is a redeemable ETH claim with no user leverage inside the protocol; its backing ratio read about 1.1014 at block 26384083.
  • Bug bounty: Immunefi lists a USD 1,000,000 maximum; the captured page shows it live since 08 July 2023 and last updated 01 January 2025, so current status is not confirmed.
  • Operator safeguards: a withdrawal-credential check before lending 28 ETH exists, per the official ETHx security explainer.
  • Incident handling (NearX, 2022): Stader published a post-mortem and a make-whole plan; completed repayment is not shown.

Contradictions & Inconsistencies

  • Timelock reach: the overview describes ETHx contracts as owned by the 7-day timelock, and the docs say it owns "all" ETHx contracts. The governance report finds exceptions: ETHxPriceOracle and the other-chain ETHx proxies upgrade through 3-signature Safes. The governance finding takes precedence because it rests on on-chain reads.
  • Manager multisig size: the official docs page says 3-of-5; a newer page and the on-chain read say 3-of-6. Use 3-of-6.
  • What the oracle proxy feeds: governance says the file does not show what depends on ETHxPriceOracle; the on-chain file shows ETHxRateProvider reads it (Risk 2). The economic report separately says a Stader oracle contract supplies the USD price to other protocols, without naming the address.
  • Insurance fund control: governance leaves a no-delay withdrawFund holder unresolved; economic describes the fund's key as a 3-of-6 Safe. The on-chain file ties it to the Manager Safe 0xAAfb…07AF (Vector 1).
  • Effect of the paused registry: the overview says effects on deposits and withdrawals are not documented; the economic report says the pause "blocks new permissionless operator registrations but does not touch user exits". The cause of the pause is unstated in all reports; treat the economic wording as an inference.
  • Oracle turnover brake: the documentation says the operator addition or removal rate is limited to one per week, but the measured cooling period for trusted-node changes is zero at block 26384083. Another enforcement mechanism is possible; treat the documented limit as unconfirmed.
  • Upgrade notice versus exit completion: governance describes 7 days of notice, and economics gives typical redemptions of 7 to 10 days. Both can be true, but notice is not a guaranteed escape window (Vector 5).
  • Bridge settings versus the overview: the overview says bridge settings were not read, while the economic report and the route file supply selected settings. Use the measurements, and keep the unread OffRamps and missing prices as gaps.
  • HBARX exit time: the economic report cites about 24 hours from an unverified publisher, set by Hedera's native unstaking time; no stronger source confirms it. (source reliability: unverified publisher)
  • Data completeness labels: the economic report's score block says data completeness 1.00 and also marks liquidity as provisional with a stated confidence of about 0.55. The code report gives 0.65. The labels do not change this analysis, but the economic evidence is weaker than "1.00" suggests.
  • Incident search: the social report records that searches for incidents after 2022 failed on a provider quota error, so only the NearX 2022 exploit is documented, and that one comes from Stader's own account. It affected an inactive product, not ETHx or HBARX. A third-party write-up of NearX (source reliability: unverified publisher) gives details the Stader account does not confirm.

Overall Risk Profile

Stader's ETHx is a conventional pooled liquid staking token with a well-timelocked core and a weaker periphery: a 3-of-6 Manager key with instant powers, a three-node oracle committee, an unquantified loss-cover fund and an unmeasured exit market. The workers' scores are 4.5/10 for code security, 4.5/10 for economic security, 4.9/10 for governance and 7.1/10 for social and team transparency. The most likely failure mode is a temporary exit disruption: a pause, safe mode or run that blocks or slows ETHx exits and opens a secondary-market discount; unpausing needs the 7-day timelock, and the safe-mode duration is not stated. The worst case is a combination of a key compromise on the Manager Safe (oracle upgrade, pause, fund withdrawal) or a correlated slashing event that exceeds the 4 ETH bond with no defined user loss rule, affecting up to the $133.0M in ETHx. A malicious core upgrade through the Community multisig is a lower-likelihood catastrophic path, because its 7-day notice is shorter than the 7 to 10-day redemption queue (Vector 5). The HBARX slice (about $23.7M) is the least assessed. The reports give no peer figures, so no ranking against other liquid staking protocols is supported here; the findings above are the basis of the call.

Data Gaps

  • Some source material could not be collected; the report may be incomplete.

  • Some protocol documentation could not be read because the site could not be reached.

  • Some protocol documentation could not be read because the page was unavailable.

  • Some protocol documentation could not be read because collection was unsuccessful.

  • Some source material is unavailable because it was not found publicly.

  • Some source material is unavailable because it could not be collected.

  • Current deployment addresses could not be confirmed for some audited contracts.

  • Source summaries cover only the first 80000 characters of 2 document(s); later content was not assessed.

  • HBARX: mechanics, contract code, admin keys, audits and the retrieved Halborn Hedera report are all missing; the unverified-publisher fee and unstake claims are unconfirmed.

  • Mainnet ETHx: depth at 2% price impact, holder concentration and the asset check were not measured; mainnet ETHx and Gnosis prices are missing for some bridge holders.

  • Oracle: who can add or remove trusted nodes, the unit of the change limit, and which contract integrators read were not traced.

  • Loss cover: insurance fund balance and any user-loss payout rule; whether the NearX reimbursement completed.

  • Signers: current Community and Manager signer identities are not matched to on-chain owners.

  • Audit-to-deployment: audited versions were not compared with deployed code, so which fixes are live could not be assessed for any combined vector.

  • Pause reasons: the paused PermissionlessNodeRegistry (Ethereum) and BNB Chain StakeManager are unexplained. The Ethereum SD pool owner 0x4483…9449 was not read.

  • Dependencies without a usable report: Chainlink CCIP (unresearched, no files); Chainlink ETH/USD, LayerZero and Hedera have no dependency-index entry; WETH backing is unresolved. The bridge-route file is complete, but its CCIP rows carry unread offramps and null verifier settings, so a failure of a route could not be bounded.

  • Residual products: the official site lists MaticX and BNBx as discontinued, and MaticX contract withdrawals run until 3 August 2029 per an unverified publisher. (source reliability: unverified publisher) No residual balances are supplied, and findings on MaticX and BnbX were checked only against default-branch snapshots.

  • Incident search after 2022: failed on a provider quota error, so none was checked beyond the supplied files.

114 citations · click markers to view evidence

Preview3 · Default / Rows