The attack surface in crypto is wider than in traditional finance — from the protocol layer to the application layer, from online wallets to hardware wallets, from exchange accounts to DeFi contracts, every link in the chain can become an attack vector. This article systematically breaks down 10 major attack methods. The conclusion up front: phishing sites, leaked seed phrases, and fake support scams are the three attack types that cause the most losses for retail users worldwide (over 70% combined), while advanced threats like supply-chain attacks and side-channel attacks mostly target institutions. For everyday account setup, register your main account through Binance Official Site, install the Official Binance App on Android, and iOS users switching to an international Apple ID can reference the iOS Installation Guide.
I. Classifying Crypto Attack Methods
Broken down by attack layer:
| Attack layer |
Main threats |
Share (by losses) |
| User-side attacks |
Phishing, social engineering, leaked seed phrases |
~35% |
| Platform-side attacks |
Exchange hacks, API vulnerabilities |
~25% |
| Protocol-side attacks |
Smart contract exploits, bridge hacks |
~30% |
| Infrastructure |
DNS hijacking, BGP routing attacks |
~5% |
| Hardware-side |
Hardware wallet supply chain + side-channel attacks |
~5% |
Total crypto losses to theft in 2025 came to roughly $2.8 billion, with the single Bybit incident alone accounting for $1.46 billion (52% of the total). Excluding Bybit, losses at the retail level were concentrated mainly in phishing and leaked seed phrases.
Q: Which attack is a retail user most likely to encounter?
Ranked by probability of loss for retail users:
- Phishing sites (30-40%): impersonating exchange/wallet domains
- Leaked seed phrases (20-25%): photographed, uploaded to the cloud, or recorded in plain text
- Fake support scams (15-20%): phone calls or Telegram messages impersonating platform support
- Token-approval phishing (10-15%): on-chain signature traps
- Airdrop scams (5-10%): signing a malicious contract while claiming an airdrop
- Clipboard hijacking (3-5%): malware swapping out a copied wallet address
An ordinary user who understands the first four categories of protection can already avoid 90%+ of losses.
II. Attack Method 1: Phishing Websites
A phishing site is a fake site impersonating a real exchange or wallet, designed to trick users into entering their login credentials or private keys.
Typical characteristics of phishing domains:
- Look-alike domains: binance-vip.com, okex-login.cn, binnance.com (an extra "n")
- Homophone domains: mybinance.com, binance365.com
- Character substitution: b1nance.com (the digit 1 replacing the letter i)
- TLD substitution: binance.cm (not .com, but .cm)
Real-world case:
In November 2024, a user searched "Binance official site" on a search engine, and the first ad result was binance-vip.com (a phishing site). After logging in, the user's account was emptied of $120,000.
Protection tips:
- Never search for the official site through a search engine (over 50% of the top ad results are phishing sites)
- Bookmark the official domain directly (binance.com / okx.com)
- Check the certificate: click the padlock icon in the browser's address bar and confirm it's issued to "Binance Holdings Limited"
- Download the official app: only from an app store or the official download page
Q: How do you verify whether a site is a phishing site?
A 5-step check:
- Check that the domain is a complete, exact match (don't just look at the prefix)
- Check the SSL certificate (click the padlock icon → issuing organization)
- Look up the domain's WHOIS record (registration date, registrant location)
- Check whether the homepage content matches (phishing sites usually have slight UI differences)
- Cross-check against official announcement channels (Twitter / Telegram)
If you find a suspicious site, close the browser immediately and do not enter any information.
III. Attack Method 2: Seed Phrase Theft
A seed phrase (also called a mnemonic phrase) is the human-readable form of a wallet's private key — 12 to 24 English words. The seed phrase is the key to every asset in that wallet.
Common ways a seed phrase gets leaked:
- Saved as a photo in the phone's gallery (synced to the cloud → iCloud or Google Photos gets compromised by a hacker)
- Uploaded to cloud storage (OneDrive, Google Drive, other cloud drives)
- Screenshotted and sent through a chat app (WeChat, QQ, etc.)
- Recorded in plain text in a notes app or spreadsheet (read by remote malware)
- Pasted into a search engine by mistake
- Picked up by a voice assistant (Siri / Google Assistant accidentally recording the phrase)
- Extracted by an OCR text-recognition tool (reading the seed phrase out of an image)
Real-world case:
In March 2025, a user saved a screenshot of their MetaMask seed phrase to iCloud. Three months later, the iCloud account was phished, the seed phrase was stolen, and $220,000 in ETH was drained from the wallet.
Protection tips:
- Write it down on paper by hand (leaves no digital trace)
- Use a metal backup (fire- and water-resistant, e.g. Cryptosteel)
- Store it in a physical safe (a home safe or a bank safe-deposit box)
- Never photograph it
- Never upload it to any cloud service
- Never send it to anyone (including "support")
Q: Is it safe to store a seed phrase in a password manager (like 1Password)?
Partially safe. Encrypted password managers like 1Password and Bitwarden are better than plain-text storage, but risks remain:
- The master password gets phished: if a 1Password master password is phished, everything inside is exposed
- The device gets remotely controlled: malware may be able to capture the data at the moment of decryption
- An official breach: 1Password has never had a breach historically, but it's theoretically possible
Recommendation: keep the primary seed-phrase backup on paper or metal, and use a password manager only for an "emergency copy," with an independent master password.
IV. Attack Method 3: Fake Support Scams
Fake support scams are one of the schemes that trip up users most often.
Typical scenarios:
- Someone adds you on a chat app out of nowhere, claiming to be "Binance support agent XXX"
- A Telegram group impersonating an official account, with a username like @binance_support_official
- A phone call impersonating support, claiming "your account shows unusual activity and needs verification"
- A remote-control trick, getting the user to install TeamViewer or AnyDesk
Common scam scripts:
- "Your account has an abnormal login and needs to be dealt with immediately"
- "There's a problem with your KYC information and it needs to be re-verified"
- "You can speed up your withdrawal, but you need to transfer in a deposit first"
- "Your account has been frozen by risk control, and you need to provide your seed phrase to unfreeze it"
Real-world case:
A user received a call from someone claiming to be "Binance support," who accurately read back the user's name, phone number, and account registration date (information that had already circulated on the black market after a 2025 Binance data leak). The scammer got the user to install AnyDesk for remote control and drained $80,000 from the Binance account within 30 minutes.
Protection tips:
- Anyone who reaches out to you first claiming to be "support" is a scammer
- Official support only responds when you contact them first (through in-app support or the official website)
- Official support never asks for your seed phrase, private key, or password
- Official support never asks you to install remote-control software
- Official support never asks you to transfer funds to any "safe account"
Q: How do you reach real official support?
The only correct ways:
- Binance: in-app support (the question-mark icon in the top right) or the live chat on binance.com
- OKX: in-app support (Profile → Help Center → Live Chat) or okx.com
- Bybit: in-app support (Profile → Live Chat)
Official support will never reach out to you first (except as a reply to a support ticket you previously opened). Any unsolicited contact is a scam.
V. Attack Method 4: Token Approval Phishing
Every new DeFi protocol interaction requires you to first "Approve" a contract to access tokens in your wallet. Malicious-contract approval phishing is the most common on-chain attack of 2024-2026.
Typical scenarios:
- Airdrop phishing sites: claiming you can claim an airdrop, but actually getting you to approve a malicious contract
- NFT minting phishing: a "free" NFT mint where the contract secretly drains your tokens
- Fake DApp sites: impersonating Uniswap or OpenSea, draining your assets after you approve
- Malicious airdropped tokens: unknown tokens sent to your wallet that trigger a theft when you interact with them
How to spot approval phishing:
- Check the contract address: MetaMask displays the contract address during signing — compare it against the official documentation
- Check the approval amount: be alarmed if it shows "Unlimited" or an absurdly large number
- Check which token is being approved: is it a valuable token in your wallet (USDC / USDT / WETH, etc.)?
Real-world case:
In Q3 2025, a user saw an "AI project airdrop" link on Twitter, connected their wallet, and signed an approval transaction that granted a malicious contract unlimited access to their USDC. Ten minutes later, $12,000 in USDC was drained from the wallet.
Protection tips:
- Only interact with official DApp sites (type the domain in manually)
- Carefully check the contract address and amount every time you approve
- Use Revoke.cash to periodically revoke old approvals
- Keep large holdings in a separate "main wallet" that never connects to any DApp
Q: How do you revoke a token approval you've already granted?
To revoke an approval:
- Go to revoke.cash
- Connect your wallet (MetaMask / WalletConnect)
- Review all approvals: which tokens, which contracts, what amounts
- Revoke any suspicious approval by clicking "Revoke" (requires a gas fee)
It's recommended to do an approval audit once a month, revoking access for any protocol you no longer use.
VI. Attack Method 5: Clipboard Hijacking
Clipboard hijacking malware is a "quiet but deadly" attack.
How it works:
- The user copies a wallet address, 0x1234...abcd, intending to paste it into a withdrawal field
- The malware detects that the clipboard content is a wallet address
- The malware swaps it for a hacker's address, 0x9876...zzzz (similar-looking characters, but actually the hacker's address)
- The user pastes it without checking the first and last few characters → the assets go to the hacker
Typical sources of this malware:
- Pirated Office / Adobe software: bundled into cracked installers
- Tools downloaded from download sites: unofficial sources
- Malicious browser extensions: unreviewed, low-profile extensions
- PDF reader vulnerabilities: opening a malicious PDF file
Real-world case:
In 2024, a user downloaded a "cracked Adobe Premiere" from a piracy download site, and malware sat undetected on the machine for 3 weeks. One day, the user tried to withdraw 5 ETH to their own MetaMask wallet — the copy-pasted address was swapped for the hacker's, and the user lost 5 ETH (about $20,000). The user only later noticed that the first character of the address had been altered.
Protection tips:
- Verify the first and last characters every time you paste an address (first 4 + last 4)
- Use a withdrawal whitelist (supported by Binance/OKX)
- Don't install pirated/cracked software
- Run a full antivirus scan (Windows Defender is sufficient)
- Test large withdrawals with a small amount first (send 0.01 ETH first to verify)
Q: How do you tell if your computer has clipboard-hijacking malware?
A simple test:
- Copy a wallet address (for example, 0x1234567890abcdef)
- Open a notes app and paste it
- Check whether the pasted content matches the original exactly
If it doesn't match, your computer very likely has clipboard-hijacking malware. Immediately:
- Run a full antivirus scan
- Check your browser extensions (remove anything suspicious)
- Reinstall the OS (the most reliable fix)
VII. Attack Method 6: Exchange Account Theft
Common paths to an exchange account getting compromised:
- Weak password: using something easy to guess (like 123456 or a birthday + name)
- Password reuse: using the same password on multiple platforms (one leak compromises everything)
- 2FA bypassed: SIM-swap attacks / SMS interception
- Leaked API keys: API keys accidentally exposed on GitHub or leaked through a third-party platform
- Device remotely controlled: a login session gets hijacked
SIM Swap Attacks:
A scammer impersonates the user and requests a replacement SIM from the carrier → receives the reissued SIM and starts intercepting the user's texts → resets the exchange login password and 2FA → withdraws the funds.
SIM-swap cases increased across 2024-2025, mainly targeting crypto influencers and public figures.
Protection tips:
- A strong password: 16+ random characters (generated with a password manager)
- A different password for every platform
- Use TOTP for 2FA rather than SMS (Google Authenticator, Authy)
- Turn on hardware 2FA for key accounts (YubiKey)
- Set an IP whitelist on API keys
- Regularly review your login history and device list
- Disable any API keys you don't need
Q: Is SMS-based 2FA really unsafe?
SMS 2FA is better than no 2FA at all, but it's far weaker than TOTP:
- SIM swaps: a scammer can intercept SMS by swapping the SIM
- SMS-interception malware: phone malware can read incoming texts
- SS7 attacks: a telecom protocol vulnerability (an advanced threat, mainly targeting institutions)
Strongly recommend enabling Google Authenticator or a YubiKey, and disabling SMS 2FA (or keeping it only as a backup recovery method).
VIII. Attack Method 7: Smart Contract Exploits
Smart contract exploits are the biggest risk in DeFi protocols.
Common vulnerability types:
| Vulnerability type |
Description |
Notable case |
| Reentrancy attack |
A contract's recursive calls cause inconsistent state |
The 2016 DAO hack |
| Flash loan attack |
Using a flash loan to manipulate price or voting |
Cream Finance, 2022 ($130 million) |
| Integer overflow |
A large-number overflow corrupts balances |
Beauty Chain (BEC), 2018 |
| Missing access control |
A critical function has no permission restriction |
Wormhole, 2022 ($320 million) |
| Oracle manipulation |
The price feed gets manipulated |
Mango Markets, 2022 ($116 million) |
| Governance attack |
A malicious proposal pushed through with governance tokens |
Tornado Cash governance hijack, 2023 |
User-level protection:
- Only use protocols with large TVL and a long track record (Aave / Curve / Uniswap)
- Avoid new protocols (< 6 months old): not yet battle-tested
- Check the audit reports: has it been audited by well-known firms like OpenZeppelin or Trail of Bits?
- Check for insurance coverage: does the protocol have coverage through Nexus Mutual or Sherlock?
- Diversify your exposure: no single protocol should exceed 30% of your total DeFi position
Q: Is an audited protocol automatically safe?
Not necessarily. An audit reduces the risk from known vulnerabilities, but:
- Audit budgets are typically $50,000-$200,000: projects may cut corners on audit depth
- New types of vulnerabilities: an audit can't catch vulnerability classes that haven't been discovered yet
- Operational risk: an audit doesn't cover the project team's operational risk (such as a rug pull)
The safest protocols have: 2+ audits from well-known firms + 2+ years of operating history + TVL over $100 million + no major historical incidents.
IX. Attack Method 8: Cross-Chain Bridge Hacks
Cross-chain bridges connect different blockchains, and were a hotbed of hacks in 2022-2023.
Attack vectors against bridges:
- Multisig private keys phished (Ronin Network, $625 million)
- Signature verification vulnerability (Wormhole, $320 million)
- Forged cross-chain messages (BNB Chain Bridge, $570 million)
- Leaked operator keys (Harmony Horizon, $100 million)
User protection:
- Bridge less, and only in small amounts (under $50,000 per transaction)
- Prefer mainstream bridges (LayerZero, the upgraded Wormhole, Connext)
- Avoid new bridges (operating for less than 12 months)
- Split large cross-chain transfers into batches (5-10 smaller transactions)
- Watch the bridge's real-time on-chain TVL: stop using it immediately if TVL drops abnormally
X. Attack Methods 9-10: Advanced Threats
Attack Method 9: DNS Hijacking
DNS hijacking refers to attackers gaining control of a name server and pointing an official domain to a phishing site.
Notable cases:
- KyberSwap's DNS was hijacked in 2022, and users had their assets stolen after connecting their wallets
- Curve Finance's domain was hijacked for 4 hours in 2024
Protection:
- Use a trusted DNS resolver (Cloudflare 1.1.1.1, Google 8.8.8.8, etc.)
- Enforce HTTPS (via HSTS headers)
- Check the certificate (be alert if it suddenly changes)
Attack Method 10: Hardware Wallet Supply-Chain Attacks
Hardware wallets (Ledger, Trezor) are generally very secure, but there are risks around purchasing and using them:
- Buying used or from a third party: the device may have been tampered with and repackaged
- Tampering in transit: the shipment gets swapped for a device with malware implanted
- Official supply-chain vulnerabilities: Ledger's 2020 customer data leak
Protection:
- Buy directly from the official site (Ledger.com / Trezor.io)
- Never buy a used or opened device
- Always generate your own seed phrase on first setup (never accept a pre-filled seed phrase)
- Check for tamper-evident packaging
XI. A Comprehensive Protection Checklist
A security checklist for ordinary users:
Basic security (everyone should do this):
- [x] Turn on 2FA on the exchange (using Google Authenticator)
- [x] Set an anti-phishing code (to block fake emails)
- [x] Set a withdrawal whitelist
- [x] Use a strong password + a password manager
- [x] Use a dedicated email with a strong password
- [x] Write the seed phrase by hand and store it physically
Intermediate protection (for $50,000+ in assets):
- [x] Use a hardware wallet (Ledger / Trezor)
- [x] Spread assets across 2-3 exchanges
- [x] Keep large amounts in cold storage
- [x] Use a YubiKey for hardware 2FA
- [x] Periodically revoke token approvals
- [x] Set an IP whitelist on API keys
Advanced protection (for $500,000+ in assets):
- [x] Use a multisig wallet (Safe Wallet)
- [x] Spread assets across 4-5 licensed exchanges
- [x] Keep most assets in hardware cold storage
- [x] Use a dedicated device for crypto trading (not shared with everyday use)
- [x] Regularly audit accounts, approvals, devices, and balances
Q: What proportion of assets should a regular user keep in hardware cold storage?
By total asset size:
| Total assets |
Recommended cold-storage share |
| Under $10,000 |
0% (unnecessary) |
| $10,000-$100,000 |
20-30% |
| $100,000-$1,000,000 |
40-60% |
| Over $1,000,000 |
60-80% |
The larger the total assets, the higher the cold-storage share should be. Truly large holdings (over $10 million) should consider multisig + cold storage + geographic distribution.
XII. Incident Response: What to Do After an Attack
1. Seed phrase leaked → move assets immediately
- Generate a new seed phrase in a new wallet
- Transfer all assets to the new wallet
- Revoke all approvals on the old wallet (via Revoke.cash)
2. Exchange account compromised → emergency steps
- Log in and disable API access
- Contact platform support to freeze the account
- File a police report and get on-chain analytics firms to help trace the funds
3. On-chain DeFi assets stolen
- On-chain transactions are public, so you can track the flow of funds
- If the funds land on a CEX, you can contact that exchange's risk team to freeze them
- For major incidents, firms like Chainalysis or TRM Labs can assist
4. Wallet compromised through a phishing signature
- Move any remaining assets to a new wallet immediately
- Revoke every approval on the compromised wallet
- Stop using the compromised wallet entirely
XIII. Frequently Asked Questions
Q: Does insurance actually pay out?
DeFi insurance (Nexus Mutual, Sherlock, etc.) only pays out for a protocol being hacked, not for user error (phishing, leaked private keys, etc.). Insurance is mainly a supplementary tool for institutional clients and isn't very useful for retail users.
Q: Can you report a scam to the police?
Yes. Law enforcement in many jurisdictions will typically open a case (crypto fraud is usually handled under fraud or illegal-fundraising statutes), but recovery rates are extremely low (under 5%). Prevention is 100 times more valuable than recovery.
Q: Is it true that wallets can be monitored?
Theoretically possible, but rare in practice. Transactions through mixers like Tornado Cash get flagged by on-chain analytics firms, but ordinary wallet activity isn't specifically targeted for surveillance. Ordinary users shouldn't overthink their privacy — solid basic security is enough.
Q: Is a phone or a computer safer for crypto transactions?
Phones are somewhat safer:
- iOS's sandboxing is stricter
- App store review is more thorough
- Malware spreads more slowly
Computers carry more risk:
- More miscellaneous software installed, more malware exposure
- Browser extensions pose a bigger risk
- More copy-paste attacks
Best practice: use your phone for everyday app usage, and a dedicated, secured computer for large transactions.
Q: Is it risky to trade through a VPN?
A VPN itself doesn't increase crypto trading risk, but:
- Free VPNs may log your traffic: choose a reputable, paid VPN (like ExpressVPN)
- Shared public IPs can get flagged: an exchange may treat a shared VPN IP as suspicious and restrict your account
Prefer a paid VPN with a dedicated IP, and avoid free or shared VPNs.
Q: Is putting everything in a hardware wallet automatically safe?
No. A hardware wallet reduces the risk of a stolen private key, but:
- A lost seed phrase still means a 100% loss of assets
- Signing traps: a hardware wallet can still be tricked into signing a malicious transaction
- Physical risk: coercion or theft
The most robust strategy: split seed-phrase storage across multiple locations + use a multisig wallet (Safe Wallet) + physically distribute holdings geographically.
Q: Is MetaMask safe?
MetaMask itself is generally secure, but user behavior carries the most risk:
- The seed phrase is stored in an encrypted local file (password strength determines its security)
- The browser extension can be called by a phishing site
- Historically, MetaMask users have been the most common target of phishing sites
Best practices for using MetaMask:
- Set a password of 12+ characters
- Don't connect to unfamiliar DApps
- Pair it with a hardware wallet (connect Ledger to MetaMask)
Q: How do you spot a project that's likely to rug pull?
Warning signs of a rug pull:
- Liquidity isn't locked (the team can pull the pool at any time)
- The token contract's admin functions haven't been renounced (the team can mint unlimited tokens)
- An anonymous team (no KYC'd team information)
- Abnormal tokenomics (the team holds 80%+ of supply)
- No audit (no third-party audit at all)
Avoid a project immediately if it shows 2 or more of these signs.
Data Sources
- Chainalysis 2025 Annual Crypto Crime Report
- TRM Labs on-chain incident database (2024-2026)
- Immunefi DeFi vulnerability reports
- SlowMist's annual review of on-chain incidents
- Official security incident announcements from various exchanges
Risk disclosure: this article is an independent third-party summary and does not constitute investment or legal advice. Crypto security is a fast-moving field and threats continue to evolve. All data is current as of April 2026.