2026 OKX Web3 Security Mid-Year Report

Jointly produced by OKX Web3 Security Team, SlowMist, and OtterSec

Preface

In the first half of 2026, if we only look at publicly reported losses, the crypto industry appears to be becoming safer, with losses down nearly 60% compared to last year. However, this decline is almost entirely due to the absence of a single, massive loss event from the same period last year, rather than a reduction in hacking activity itself. In fact, the opposite is true: the frequency of attacks is increasing, techniques are evolving, and attackers' targets are shifting from "code" to "people."

Many projects have historically prioritized contract audits and on-chain activity monitoring. Yet, in the first half of this year, the biggest losses often didn't stem from the contracts themselves. They could be a blindly signed administrative transaction, a leaked cloud key, a developer's computer infected with malware, or even a convincingly real-looking video conference.

For OKX, security isn't a slogan in a report; it's a real risk dealt with daily: malicious addresses, phishing domains, high-risk signatures, abnormal approvals, cross-chain risks, scam tokens, and social engineering attacks are all persistent challenges. This report aims to do two things: first, clearly outline the noteworthy changes in attack patterns from the first half of the year; second, share some of OKX's practices in wallets, risk control, cross-chain operations, and Agent security as a starting point for industry discussion.

Chapter 1 · Security Landscape in the First Half of 2026: Scattered Events, Concentrated Losses

Declining Losses Do Not Equal Fewer Attacks

A phenomenon easily misinterpreted in the first half of this year is the decline in total stolen funds. However, there have actually been more security incidents. According to incomplete statistics from SlowMist's Hacked archive, there were 182 publicly reported security incidents in the first half of the year, resulting in losses of approximately $956 million. Compared to 121 incidents and approximately $2.373 billion in losses during the same period in 2025, the number of incidents increased by about 50% year-over-year, while total financial losses decreased by about 60% year-over-year. The apparent decline in losses is mainly because the previous year had a single, extremely large event. The absence of a similarly sized extreme event this year does not mean attackers have stopped. After removing anomalies, comparable losses this year have actually increased. Attack activity hasn't weakened; it has shifted towards higher-frequency, more dispersed strikes.

For comparison, the largest publicly disclosed security incidents in the first half of 2025 were:

[Table]

First, large-scale losses are increasingly occurring outside of contract code. The most severe losses in the first half of the year didn't come from vulnerabilities like smart contract reentrancy or precision errors, but from operational-level breaches – induced blind signing of administrative transactions, contaminated single validator nodes, and stolen cloud signing keys. In terms of incident count, contract and logic vulnerabilities remain the primary cause (85 incidents). However, in terms of loss amounts, supply chain attacks topped the list at approximately $298 million, followed by contract vulnerabilities (approximately $152 million) and private key leaks (approximately $130 million). This clearly illustrates that for any project, relying solely on an audit report is insufficient. Even if a contract has no obvious on-chain logic flaws, attackers can bypass the contract itself and strike at the weakest link if there are single points of risk in the signing process, cloud keys, cross-chain verification, or operational systems. The boundary of security has long expanded from "is the code secure?" to "who can sign, where are keys stored, does verification rely on a single point, and is the operation trustworthy?"

Second, ordinary users are becoming the primary target of attacks. As the cost of attacking protocols rises, attackers shift their focus to users. Phishing websites, malicious browser extensions, search engine ad poisoning, fake customer service, fake job offers, malicious meeting software, clipboard hijacking, and fake 2FA verification are all user-side attack methods that have repeatedly appeared in the first half of the year. These attacks don't necessarily rely on sophisticated technology; they exploit users' habits and trust in daily operations – trusting the first search result, trusting high app store ratings, trusting a link from a friend's account, trusting a "real person" in a video conference, trusting a page that says "Security Verification." AI makes all of this cheaper and more realistic: attackers can batch-generate phishing content, forge identities, clone voices, and create deepfake videos, polishing once crude scams into ones convincing enough to fool experienced users. The success of an attack depends less and less on whether the user "understands the technology," and more on whether the attacker can precisely hit the user's blind spots of trust.

Third, while thriving, AI Agents are also gradually becoming new prey for hackers. As Agents evolve from being able to "chat" to being able to "act" – calling tools, reading context, controlling assets, initiating transactions – the greater their capabilities, the more severe the consequences of a compromise. In the past, prompt injection was more of a concept in the field of model security, where the worst outcome might be getting the model to say something it shouldn't. However, when an Agent can sign transactions and transfer funds, a malicious instruction disguised as normal input can directly translate into a real on-chain loss. Attacks targeting this "perception-to-action" chain are becoming a threat as dangerous as attacking private keys.

Chapter 2 · Project Perspective: The Biggest Losses for Projects Often Occur Outside of Contracts

For project teams, the key takeaway from the first half of 2026 is: problems in operational aspects can be just as severe, or even more severe, than contract vulnerabilities.

Below are the major project security incidents of the first half of 2026, compiled by OtterSec based on loss amounts. They cover several common high-value attack surfaces: cross-chain bridges, verification infrastructure, signing processes, cloud keys, developer devices, oracles, and access controls.

[Table]

KelpDAO: Attacker Broke the Verification Path, Not the Contract

KelpDAO was the single largest loss event in the first half of the year. The most alarming aspect is that the attacker didn't directly attack the contract logic but instead attacked the verification path of cross-chain messages. According to OtterSec's post-mortem, the attacker compromised LayerZero's internal RPC nodes while launching a DDoS attack against honest external nodes. Consequently, the sole DVN relied upon by the bridge primarily saw forged data from the attacker and authorized a withdrawal without a corresponding real burn. Approximately 116,500 rsETH were transferred out, of which around $75 million was later frozen.

A single DVN configuration has long been considered high-risk. In the past, this was largely a theoretical risk discussed in architecture reviews. After the KelpDAO incident, it became a nearly $300 million real-world loss. The direct reminder for project teams is: do not let a single validator, a single RPC, a single price feed, or a single signing path determine whether funds can leave. As long as there's only one verification point on a critical path, attackers will prioritize studying it.

Cross-chain bridges and oracles especially need multi-source verification, redundant nodes, withdrawal rate limits, anomaly monitoring, and emergency pause mechanisms. Otherwise, even if the contracts themselves have no obvious vulnerabilities, attackers can still drain funds through off-chain infrastructure.

Drift: The Attacker Waited Six Months for a Signature

If KelpDAO exposed the fragility of infrastructure, Drift exposed the fragility of the signing process. This wasn't spontaneous phishing but a long-term, orchestrated social engineering attack. The attacker spent about six months building a relationship, waiting for a multisig signer to sign a critical administrative transaction without fully understanding its implications. The most concerning aspect of this incident was the abuse of the Solana durable-nonce mechanism. The attacker induced relevant parties to pre-sign multisig authorization transactions. At the time of signing, these transactions appeared to have no immediate effect. However, after Drift subsequently adjusted the multisig threshold, the attacker broadcast the already-obtained signed transactions, draining over 50% of the TVL in a very short time.

This type of attack reminds project teams: just because a signature appears to have "no effect" at the moment of signing doesn't mean it won't be exploited in the future. Blind signing, pre-signing, and unreadable administrative transactions should all be considered high-risk operations. Critical transactions must be clearly parseable, simulatable, and reviewable. For multisig teams, the signing process itself should be protected as a core asset.

OKX Practice: From Transaction Parsing to Signature Risk Protection

From OKX's experience, protecting against signature risks cannot stop at simply verifying "whether the transaction was initiated by the user." More critical is whether the user truly understands the consequences of a transaction before signing. Targeting high-risk scenarios like Solana durable nonce, account ownership changes, and nonce account initialization, OKX has established capabilities for identification, warning, interception, and isolation across multiple risk rules including solana_assign_account_owner, solana_init_nonce_account, and nonce_account_risk. In the first half of 2026 alone, these measures intercepted or alerted on related high-risk operations over 4 million times, protecting assets totaling approximately $526 million, helping users identify and avoid such risks before signing.

Simultaneously, OKX continues to advance its transaction parsing capabilities. Our goal is to make on-chain transactions more transparent and readable, striving for the principle of "what you sign is what you get": users should see not a string of incomprehensible calldata or instructions, but what the transaction truly intends to do, which assets it will affect, what permissions it grants, and whether there are any abnormal risks. To date, OKX has parsed and matched over 50,000 on-chain methods, helping users better understand the transactions they are signing.

Chapter 3 · User Perspective: Trust Has Become the Sharpest Weapon

User-side risks continued to rise in the first half of the year. Many attacks no longer start with a strange link but begin at entry points users are familiar with and let their guard down around: app stores, search results, friend accounts, meeting software, recruitment processes, and customer service emails.

Phishing Borrows the Shell of Real Platforms

Phishing remains the primary method for stealing user assets, but its form is evolving.

A common method is malicious browser extensions. Attackers often impersonate well-known wallet tools, copying brand names, icons, and page copy, then using fake reviews and downloads to make the extension seem legitimate. Users seeing it in the official app store are less likely to be suspicious, thinking it's the plugin they know. These extensions often use a "locally benign, remotely poisonous" strategy, meaning they don't contain malicious logic directly, making it easier to pass the store's static review. The truly dangerous phishing pages are served in real-time from a remote server. Attackers can change pages, domains, and even tailor content for different users at any time. Once you enter your seed phrase or private key, control of your assets is lost.

Another category is search engine advertising phishing. Attackers bid on keywords for popular projects, placing fake official websites at the top of search results. In a typical case from the first half of the year, a user searched for a development tool on a new computer, clicked the top ad, and then followed on-screen prompts to execute an "installation command" in their terminal. This command actually deployed clipboard-hijacking malware, allowing the attacker to alter page content. Later, when the user tried to transfer about $20,000, the recipient address was automatically replaced, sending the funds to the attacker. These attacks are difficult to defend against because the user isn't doing anything obviously dangerous. They are just searching for an official website, downloading a tool, and copying commands – routine actions in daily work.

Social Engineering: The Most Active and Destructive

The key to social engineering isn't technology; it's getting the victim to let their guard down at the critical moment.

The most common method is impersonating someone the victim already knows. In one real case, a victim received an event invitation from a long-trusted friend. The sender insisted on downloading a specific meeting software. The victim hesitated but ultimately installed it because they trusted the friend. Hours later, their wallet was drained. It was later discovered that the friend's account had been compromised. High-influence KOLs are also heavily targeted. Attackers create high-simulation copycat accounts, using the public figure's reputation to lure fans into fake events, fake airdrops, or fake investment groups. For ordinary users, the difficulty isn't judging if a stranger is trustworthy, but knowing whether an account that looks like a friend or celebrity has been compromised or impersonated.

Recruitment and interview scams are also becoming more targeted. Attackers first contact victims through "technical interviews," "operations interviews," or "volunteer interviews," asking them to share their screen, open their wallet, and demonstrate DeFi experience. On the surface, this seems like part of the interview process; in reality, the attacker is recording wallet addresses, holdings, frequently used protocols, and operational habits. In one real case observed, the attacker learned about the victim's recent interacting protocols and preferences during a fake interview. They then forged a highly customized airdrop page based on a protocol the victim actually used, sending tailored phishing messages that ultimately defrauded them of about $88,000.

Two New Techniques to Watch Out For

First is the fake "2FA Security Verification" scam.

Attackers send emails disguised as official wallet communications, using lookalike domains that differ by just one character, combined with countdown timers to create urgency, prompting users to enter their seed phrase for "verification." It must be reiterated: any page requesting your seed phrase for verification, authentication, recovery, or upgrade is a scam. A seed phrase is not a verification code; it is the control key to your assets. No legitimate wallet will ever ask for it via a webpage for any reason.

Second is business process fraud.

These attacks don't look like phishing; they look like normal work. Attackers use business scenarios like "company legal name confirmation," "external audit," "token vesting confirmation," or "partner information supplement" as bait, sending malicious attachments disguised as Word files, PDFs, or collaborative documents. Once opened, the malware disguises itself as a system update, prompting the user to enter their system password and requesting permissions for the camera, screen recording, and keyboard logging. The target of these attacks isn't always just a personal wallet. Often, the attacker's true goal is access to the work terminal, browser sessions, password manager, cloud service permissions, and the project's internal systems.

OKX Practice: Pushing User Protection Forward to Devices and Access Points

Looking at the evolution of user-side attacks, on-chain protection alone is no longer sufficient. Many losses don't start with an on-chain transaction but much earlier: a user downloads a malicious app, installs a disguised plugin, clicks a phishing website, or continues signing and transferring funds on a compromised device. Therefore, OKX is shifting user protection upstream from on-chain transactions to devices, applications, and access points. OKX has launched a Security Scan Assistant to help users identify risky applications hidden on their devices, mitigating the risk of asset loss from malware, disguised apps, remote access tools, or clipboard hijackers. As of the publication date, OKX has completed over 200,000 risk detections, identified over 60,000 high-risk applications, and guided users to uninstall or handle them. It currently supports Windows and Android systems, with a macOS version coming soon.

Additionally, for phishing websites and malicious DApps, OKX performs risk identification and alerts users during critical scenarios like visiting suspicious URLs, connecting wallets, or initiating interactions. As of the publication date, OKX has intercepted over 7 million risky website visits, helping users avoid phishing risks before they enter seed phrases, connect wallets, or sign transactions. For users, the best security alert isn't a notification after the loss occurs, but an extra layer of interception before the risk truly lands on-chain. OKX aims to block more attacks before signing and transferring through device risk detection, URL risk identification, DApp risk warnings, and on-chain transaction parsing.

Chapter 4 · How AI is Reshaping Attacks: From Content Forgery to "Synthetic Reality"

AI's impact on the security landscape is the most profound structural change of the first half of the year. Its influence is multifaceted. On one hand, AI makes it easier for attackers to generate phishing emails, fake websites, customer service scripts, and malicious code. On the other hand, it makes forging voices, videos, identities, and community environments easier. Furthermore, when an AI Agent starts holding funds or calling transaction tools, it becomes a new target itself.

New Battlefield: When Agents Get Close to Funds and Transaction Execution

From OKX's perspective, the AI Agent is not just a new source of risk but also a potential crucial gateway for Web3 to reach more users. OKX is advancing its Agentic Wallet, allowing users to understand strategies, manage on-chain operations, and complete complex DeFi interactions through an Agent, subject to authorization and confirmation.

By integrating curated DeFi project plugins, the Agentic Wallet can consolidate multi-step on-chain operations like swaps, lending, yield management, and cross-chain transfers into a more natural interaction flow. For many ordinary users,