Zilliqa

SubQuery Network

Cosmos

Token

<p>The Cosmos EVM incident shows why closing an exploit route is not the same as restoring holders’ prior economic position. Attackers exchanged about $2.87 million of stolen assets on decentralised exchanges and sold an estimated $2.85 million through centralised exchanges, according to the <a href="https://github.com/cosmos/security/blob/main/communications/cosmos_evm_GHSA-7g4w-cg88-2cq2_post_mortem.md" rel="noopener noreferrer nofollow">Cosmos Security post-mortem</a>.</p>
<p>By the time the affected software was patched, those transactions could not be reversed and the token’s prior liquidity conditions could not be restored. A patch may therefore succeed technically while holders still face sell-side flow, pooled-staking losses, exposure beyond realised theft, or the burden of moving to a replacement environment. That distinction does not establish that every exploit causes a measurable price move or that this patch failed.</p>
<h2 id="cosmos-evm-s-flaw-released-vested-tokens-rather-than-minting-new-ones">Cosmos EVM’s flaw released vested tokens rather than minting new ones</h2>
<p>The Cosmos EVM vulnerability affected production chains running versions below v0.6.2 or v0.7.2. It arose from inconsistent token-balance accounting between Cosmos EVM and the Cosmos SDK, allowing attackers to extract legitimate tokens from vesting accounts without increasing total token supply.</p>
<p>That last point matters. A supply-creation bug and an accounting exploit involving restricted, already-issued tokens are not the same event. No additional units need to be minted for holders to confront an effective change in the assets available for sale. Tokens subject to <a href="https://cryptodaily.co.uk/glossary/understanding-the-importance-and-types-of-token-lockup-in-crypto" rel="noopener noreferrer nofollow">vesting</a> are, by design, not meant to have the same immediate market availability as unrestricted balances. If they can be extracted and traded early, the relevant economic change is in accessible supply, not necessarily headline supply.</p>
<p>The post-mortem identified six affected networks. Its estimates of roughly $2.87 million exchanged on DEXs and $2.85 million sold through CEXs put a concrete scale on the route from an accounting discrepancy to market activity. For token holders, the concern is less whether the protocol’s maximum or total supply fields changed than whether assets that should have remained constrained became available to counterparties across trading venues.</p>
<p>It also explains why “no new tokens were minted” can be an incomplete reassurance. The phrase correctly describes one limit of the incident. It does not establish that the timing of circulation was unchanged, that victims were made whole, or that liquidity was unaffected by assets released from accounts intended to vest over time.</p>
<h2 id="public-disclosure-turned-patch-deployment-into-a-race-across-six-networks">Public disclosure turned patch deployment into a race across six networks</h2>
<p>A fix in a repository does not protect every production chain using the affected code. Each operator must identify its exposure and deploy a protected version; across six networks, practical protection depended on that execution as well as on corrected code.</p>
<p>The Cosmos post-mortem says a public pull request disclosed the vulnerability and a detailed exploitation path before the first known incident, although maintainers had prepared a fix. The disclosure left affected production chains facing a deployment task after the route had become visible.</p>
<p>For holders, “patched” can therefore describe four distinct events: discovery of the flaw, availability of corrected software, deployment by affected chains, and containment of already-extracted assets. Those events can occur at different times, and a patch announcement alone does not establish which chains deployed the fix, whether funds were extracted, whether attackers converted proceeds, or whether realized losses and subsequent market sales were addressed.</p>
<h2 id="a-patched-exploit-can-leave-holders-with-losses-diluted-staking-pools-or-migration-work">A patched exploit can leave holders with losses, diluted staking pools or migration work</h2>
<p>SubQuery Network reported that five transactions drained 382,433,441 SQT tokens, worth approximately $134,000 at the time, from pooled staking balances, 272 individual staker and delegator wallets, deployment boosters and the treasury. The project said it restored contract addresses and added onlyOwner controls, according to its <a href="https://subquery.network/blog/subquery-network-security-incident-report" rel="noopener noreferrer nofollow">incident report</a>. The fix addressed the disclosed access-control weakness, but the listed pooled balances and participant wallets remained among the sources drained.</p>
<p>The incident illustrates why residual costs may not appear as an immediately tradeable balance. Delegators can be exposed through shared contracts and pools, and treasury stakeholders through resources intended for development or operations, so the consequences can be distributed across collective infrastructure.</p>
<p>Zilliqa reported 6,772 exposed accounts, but said 51 were known to have been drained, involving 683,130,969.66 ZIL in proven theft. It decided to retire the legacy environment and require migration to Zilliqa EVM, according to its <a href="https://www.zilliqa.com/ledger-incident/" rel="noopener noreferrer nofollow">incident status page</a>. Holders who were not drained may still face access, compatibility and user-action burdens during that transition.</p>
<h2 id="rapid-containment-does-not-establish-that-the-security-surface-is-closed">Rapid containment does not establish that the security surface is closed</h2>
<p>The speed of a response remains important. Hyperbridge said an attacker forged a proof using an out-of-bounds leaf and drained its Token Gateway. The gateway was paused within hours, and a permanent patch was deployed in under 72 hours.</p>
<p>The containment steps addressed the specific route used in the incident, but the same <a href="https://blog.hyperbridge.network/april-13-post-mortem/" rel="noopener noreferrer nofollow">Hyperbridge post-mortem</a> said follow-up audits identified 14 additional vulnerabilities, including one critical issue. Those findings do not establish that the issues were exploited; they do show why a successful response to one observed attack should not automatically be read as a complete assessment of the surrounding security surface.</p>
<p>For token holders, that difference affects how they interpret recovery narratives. A paused gateway can halt a drain. A permanent patch can remove the identified weakness. Subsequent audit findings may nevertheless require further upgrades, governance decisions or operational changes before confidence in the wider system can be reassessed.</p>
<p>There is no single holder-impact metric that captures this. An exploit’s direct loss, the market treatment of extracted assets, the dependence of stakers on pooled contracts, and the quality of post-incident review all describe different parts of the exposure. A narrow technical question—was the bug fixed?—therefore cannot carry the whole economic analysis.</p>
<h2 id="bug-bounties-price-the-value-of-prevention-against-residual-holder-costs">Bug bounties price the value of prevention against residual holder costs</h2>
<p>The value of finding a flaw before exploitation is clearest when set against the costs that cannot be cleanly patched afterwards. Ethereum’s <a href="https://cryptodaily.co.uk/glossary/understanding-bounties-in-the-cryptocurrency-world-a-complete-guide" rel="noopener noreferrer nofollow">bug-bounty</a> programme covers execution- and consensus-layer client bugs, including specification non-compliance, denial-of-service vulnerabilities and issues capable of causing irreparable consensus splits. It offers rewards of up to $1 million.</p>
<p>That ceiling does not place a universal price on every blockchain vulnerability. It does indicate the economic importance Ethereum assigns to reporting severe defects before they can cause irreversible disruption. A bounty payment is comparatively contained: it can avoid stolen assets entering DEX or CEX markets, pooled stakes being drained, or users being required to move from a retired environment.</p>
<p>The Cosmos case supplies the most direct comparison. Once assets extracted from vesting accounts had been exchanged or sold, a corrected release could prevent repetition of the accounting exploit but could not unwind the reported trading activity. The same basic asymmetry applies to the SubQuery and Zilliqa examples: controls can be added and environments can be retired, but the remedial work begins after holders, delegates, treasuries or users have already absorbed some form of risk.</p>
<p><a href="https://cryptodaily.co.uk/tag/security" rel="noopener noreferrer nofollow">Security</a> spending is therefore not only about preventing a protocol from going offline. It is also an attempt to preserve the conditions around a token that code changes cannot recreate after the fact: scheduled restrictions on assets, custody integrity, stable participation arrangements and the ability for holders to remain in an ecosystem without a forced recovery process.</p>
<p>Ethereum’s <a href="https://ethereum.org/bug-bounty/" rel="noopener noreferrer nofollow">bug-bounty programme</a> frames that preventive logic explicitly through rewards of up to $1 million. The Cosmos post-mortem shows the other side of the equation: even where total supply does not increase, assets released from vesting accounts can still be exchanged and sold before the patch has finished doing its work.</p>
<p>Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.</p>

Why a Patched Blockchain Vulnerability Can Still Matter to Token Holders

Table of Contents

Crypto Daily™

Other articles published on Aug 30, 2026

Join our free newsletter for daily crypto updates!