Permit and Permit2 Phishing: Signature-Based Attacks Explained

Definition

Permit and Permit2 phishing represent a sophisticated category of cyberattacks in the decentralized finance (DeFi) ecosystem, specifically targeting token approval mechanisms. Unlike traditional phishing scams that aim to steal private keys or trick users into sending tokens directly, these attacks manipulate a user's trust to obtain a cryptographic signature. This signature, often an EIP-712 structured message, grants an attacker the authority to transfer specific tokens from the victim's wallet without requiring a visible on-chain approve() transaction from the victim's address. The core innovation exploited here is the ability for users to pre-authorize token spending through off-chain signatures, a feature designed to enhance user experience and reduce gas fees.

Permit and Permit2 phishing are signature-based attacks where malicious actors trick users into signing off-chain messages that grant the attacker permission to spend the user's tokens, bypassing traditional on-chain approval visibility.

Key Takeaway

The fundamental danger of Permit and Permit2 phishing lies in its stealth and the irreversible nature of the authorization. Users are accustomed to reviewing and confirming on-chain transactions, which typically involve a clear display of the asset, amount, and recipient. However, with signature-based approvals, the malicious intent is hidden within the data structure of an off-chain message. Once a user signs such a message, they have effectively given a "blank check" or a specific power of attorney to the attacker for their tokens. This authorization is not immediately visible as a transaction on the blockchain from the victim's perspective, making detection difficult until assets are already drained. The key takeaway is that signing a message in Web3 can be as impactful and dangerous as sending a transaction, and often with less transparent immediate feedback.

Mechanics

Understanding the mechanics of Permit and Permit2 phishing requires a brief overview of token approvals. Traditionally, for a decentralized application (dApp) to spend ERC-20 tokens on behalf of a user (e.g., for swapping on a DEX or depositing into a lending pool), the user first needs to send an approve() transaction to the token contract. This transaction grants the dApp (the "spender") an allowance to move a specified amount of tokens from the user's wallet. This process is on-chain, visible, and requires gas.

Permit (ERC-2612) was introduced to streamline this. It allows users to approve token spending without paying gas fees by signing an off-chain message. This message, once signed, can then be submitted on-chain by the dApp (or anyone) to update the token's allowance. The permit() function on the token contract verifies the signature and updates the allowance. This means a user only pays gas once for the actual token transfer, not for the approval.

Permit2, developed by Uniswap Labs, takes this concept further. Instead of granting approval to individual dApps for specific tokens, Permit2 acts as a universal token approval contract. A user first grants a single, long-term approval to the Permit2 contract itself. Subsequently, any dApp that integrates with Permit2 can request off-chain signatures from the user to authorize specific token transfers through the Permit2 contract. This eliminates the need for repeated approve() transactions for every new dApp or interaction, significantly improving user experience and reducing gas costs across the DeFi ecosystem. The off-chain signature, typically formatted according to EIP-712, specifies details like the token, amount, recipient, and an expiry time.

The phishing attack exploits this mechanism. A malicious website or dApp, often a convincing replica of a legitimate service, prompts the user to sign an EIP-712 message. This message, disguised as a routine interaction (e.g., "confirming login," "claiming airdrop," or "setting up a trade"), actually contains a permit or permit2 call that grants the attacker permission to transfer the user's tokens. Because the interaction is a signature request rather than a transaction, many users might not scrutinize the payload as carefully. Once the signature is obtained, the attacker can then submit this signed message to the blockchain, executing the permit or transferFrom operation through the Permit2 contract, and draining the victim's assets. The key here is that the authorization record is not a direct approve() transaction from the victim's address, but rather an execution initiated by the attacker using the victim's signature.

Trading Relevance

For active traders and participants in the DeFi space, Permit and Permit2 phishing poses a significant and often underestimated threat. Traders frequently interact with numerous dApps, including decentralized exchanges (DEXs), lending platforms, yield aggregators, and NFT marketplaces. Each interaction often involves token approvals. The convenience offered by Permit and Permit2, such as gasless approvals and streamlined workflows, makes these mechanisms attractive but also creates new attack vectors.

Traders, especially those operating under time pressure or seeking to capitalize on fleeting opportunities, might be more susceptible to quickly signing requests without thorough examination. A moment of distraction or a misinterpretation of a signature request on a seemingly legitimate platform can lead to catastrophic losses. The ability of attackers to drain funds without a traditional approve() transaction appearing on the victim's Etherscan history means that the compromise might not be immediately apparent, allowing attackers more time to move stolen assets. This type of phishing can target any ERC-20 token, including stablecoins, governance tokens, and highly liquid assets, making it a direct threat to a trader's entire portfolio. Vigilance and a deep understanding of what is being signed are paramount for anyone actively engaging with DeFi protocols.

Risks

The risks associated with Permit and Permit2 phishing are multi-faceted and severe, extending beyond simple financial loss.

One of the most immediate and devastating risks is the irreversible loss of assets. Once a malicious signature is obtained and used, the attacker can transfer tokens out of the victim's wallet rapidly. Unlike some other scams where transactions might be reversible under specific circumstances, a valid cryptographic signature for a permit or transferFrom call is an undeniable authorization on the blockchain. The speed at which funds can be drained means victims often have little to no time to react or mitigate the damage.

Another significant risk is the stealthy nature of the attack. Traditional phishing often involves fake websites asking for seed phrases or direct token transfers. Permit and Permit2 phishing, however, operates by requesting a signature for an off-chain message. This signature does not immediately trigger an on-chain transaction from the user's wallet, making it less conspicuous. The authorization is only visible on-chain when the attacker uses the signature to execute the permit or transferFrom function, often from their own address. This lack of immediate on-chain visibility from the victim's perspective can delay detection, giving attackers a critical window to move funds.

Furthermore, these attacks are highly sophisticated and difficult to detect. Phishing sites are often meticulously crafted to mimic legitimate dApps, sometimes even using similar URLs or compromised legitimate frontends. The EIP-712 signature prompts, while containing structured data, can be complex for an average user to parse and verify, especially when presented under pressure or with misleading context. The attacker's ability to target specific tokens or even grant themselves unlimited spending approval for a particular token adds another layer of danger. The broad adoption of Permit2 by major DeFi protocols means that a vast number of users are potentially exposed to this attack vector, making it a widespread concern for wallet security.

History and Examples

The evolution of phishing attacks in Web3 has seen a shift from simpler methods, like tricking users into revealing seed phrases or approving infinite allowances via approve(), to more advanced signature-based exploits. While the concept of signing off-chain messages for various purposes has existed for some time, the widespread adoption of standards like ERC-2612 (Permit) and particularly Uniswap's Permit2 has created a fertile ground for new forms of phishing.

The year 2023 and early 2024 saw a significant surge in signature-based phishing attacks. Reports indicate that phishing attacks in 2023 resulted in over $295 million in losses, and the trend continued to accelerate, with substantial amounts lost in the first half of 2024 alone. Many of these incidents involved attackers leveraging malicious Permit or Permit2 signatures. For instance, users interacting with seemingly legitimate airdrop claims, fake token swaps, or even compromised frontends of popular dApps were prompted to sign messages. These messages, unbeknownst to the victims, contained authorizations that allowed attackers to drain their wallets.

A common scenario involves a user visiting a phishing site that looks identical to a well-known DEX. The site might prompt them to "connect wallet" and then "sign a message to verify." This "verification" message is, in reality, a malicious Permit2 signature request. Once signed, the attacker gains the ability to transfer the user's tokens. The victim might only realize the compromise hours or days later when they notice their wallet balance is empty, and upon checking transaction history, they find no direct approve() transaction from their address, only the subsequent transferFrom initiated by the attacker. These attacks highlight the continuous cat-and-mouse game between security researchers and malicious actors in the rapidly evolving crypto landscape.

Common Misunderstandings

Several common misunderstandings contribute to users falling victim to Permit and Permit2 phishing. Dispelling these myths is essential for robust wallet security.

One prevalent misconception is, "I didn't send a transaction, so my funds are safe." This is fundamentally incorrect in the context of signature-based approvals. In Web3, signing a message, especially an EIP-712 structured message, can be a powerful authorization. It's not a transaction in the traditional sense of moving funds directly, but it grants another entity the permission to move your funds. This distinction is often lost on users who are primarily focused on on-chain transaction confirmations. A signed Permit or Permit2 message is a valid cryptographic proof that you authorized a specific action, and an attacker can use this proof to execute the action on your behalf.

Another misunderstanding is, "My hardware wallet or secure software wallet protects me from all phishing." While hardware wallets offer superior protection against private key theft, they do not inherently prevent a user from signing a malicious message. If a user is tricked into confirming a malicious EIP-712 signature request on their hardware wallet, the wallet will dutifully sign it. The security of the wallet lies in protecting the private key, but it cannot discern the intent behind the data being signed. Users must still carefully review the details presented on their hardware wallet screen, which can be challenging for complex EIP-712 payloads.

Finally, many users believe, "I can easily see all my token approvals on Etherscan or similar block explorers." While traditional ERC-20 approve() allowances are clearly visible as transactions from your address, Permit and Permit2 approvals obtained via off-chain signatures are not. The permit or transferFrom call that actually moves the tokens is initiated by the attacker using your signature, meaning the transaction will appear as an interaction from the attacker's address, not necessarily a direct approval from yours. This makes it harder for victims to trace the initial compromise or even realize that an approval was granted until their funds are gone. Specialized tools or a deeper understanding of Permit2's mechanics are often required to audit these types of authorizations.

Summary

Permit and Permit2 phishing represent a significant evolution in Web3 security threats, moving beyond simple private key theft or direct token transfers to exploit sophisticated signature-based approval mechanisms. These attacks leverage the convenience of gasless, off-chain authorizations, primarily through the ERC-2612 Permit standard and Uniswap's universal Permit2 contract, to trick users into signing malicious EIP-712 messages. Once signed, these messages grant attackers the power to transfer tokens from the victim's wallet without the need for a visible on-chain approve() transaction from the victim's address, making detection challenging. The risks are substantial, including irreversible asset loss, stealthy execution, and the difficulty of identifying the initial compromise. For active DeFi participants, understanding the mechanics of these signature-based approvals and exercising extreme caution when prompted to sign any message is paramount. Always verify the source and content of signature requests, and remember that a signed message can be as potent as an on-chain transaction in authorizing the movement of your digital assets.