Replay Attack
securityFull Name: Credential Replay Attack
Definition
A replay attack is a type of network security attack in which an adversary intercepts a valid data transmission -- such as a credential presentation, authentication token, or signed message -- and fraudulently retransmits or delays it to gain unauthorized access or impersonate the legitimate sender. In the EUDI Wallet ecosystem, replay attacks represent a significant threat because a successfully replayed credential presentation could allow an attacker to assume another person's digital identity, access their government services, sign documents in their name, or bypass age and identity verification checks. The EUDI Wallet Architecture and Reference Framework mandates multiple layers of replay prevention, including cryptographic nonces, session binding, timestamps, and hardware-backed device authentication, ensuring that each credential presentation is uniquely tied to a single verification transaction and cannot be reused.
How Replay Attacks Threaten Digital Identity
In traditional paper-based identity systems, replay attacks are inherently limited by the physical nature of documents -- you cannot easily copy and present someone else's physical passport without detection. However, in digital identity systems, data transmissions can be perfectly copied and retransmitted. An attacker positioned on a network (through a compromised Wi-Fi hotspot, for example) could capture the data packets containing a credential presentation and later retransmit them to a different verifier to impersonate the victim.
The EUDI Wallet faces replay attack vectors across multiple channels. In online verification scenarios (OpenID4VP over HTTPS), an attacker might capture the verifiable presentation response and attempt to replay it to a different relying party. In proximity verification scenarios (ISO 18013-5 over BLE), an attacker with Bluetooth monitoring equipment could capture the credential exchange and attempt to replay it at a later time or location. In credential issuance flows (OpenID4VCI), an attacker could attempt to replay an authorization code to obtain duplicate credentials.
The consequences of a successful replay attack on EUDI Wallet credentials are severe: identity theft, unauthorized access to government services, fraudulent financial transactions, or illegal border crossings. This is why the EUDI Wallet specification dedicates significant attention to anti-replay mechanisms and requires multiple independent protections operating simultaneously.
Anti-Replay Mechanisms in the EUDI Wallet
The primary anti-replay mechanism is the cryptographic nonce. When a verifier initiates a credential request, it generates a unique random nonce and includes it in the request. The wallet must include this nonce in its signed response (the verifiable presentation). When the verifier receives the response, it checks that the nonce matches the one it sent. Because each session uses a different nonce, a presentation captured from one session will contain the wrong nonce for any other session and will be rejected. The nonce is typically a 128-bit or 256-bit random value, making it practically impossible to predict.
Session binding provides a second layer of protection. In the ISO 18013-5 proximity flow, the device engagement process establishes ephemeral key pairs for both the wallet and the verifier. The session encryption keys derived from these ephemeral keys are unique to each interaction. Even if an attacker captures the encrypted credential data, they cannot decrypt it or retransmit it meaningfully because the decryption keys exist only for the original session and are destroyed afterward.
Device-bound key proofs (also called holder binding or key binding) provide a third layer. When the wallet presents a credential, it signs the presentation with a private key stored in the device's secure enclave. This signature proves that the presentation originated from the specific physical device that holds the credential. An attacker who captures the presentation data cannot produce this signature because they do not possess the device-bound private key. This mechanism defeats both replay and relay attacks.
Timestamps and expiration windows provide a fourth layer. Each presentation includes a creation timestamp, and verifiers reject presentations that are older than a configured threshold (typically seconds to minutes). This limits the window of opportunity for any replay attempt, even if the other mechanisms were somehow bypassed.
Real-World Replay Attack Scenarios and Countermeasures
Consider a scenario where an attacker sets up a rogue Bluetooth monitoring device near a bar entrance where patrons verify their age using EUDI Wallets. The attacker captures the BLE packets from a legitimate age verification. Later, an underage accomplice approaches the same bar. Even if the attacker could retransmit the captured packets, the verification would fail because: the verifier has issued a new nonce, the ephemeral session keys are different, and the device-bound signature cannot be reproduced without the original device.
In an online scenario, an attacker might intercept a verifiable presentation sent from a wallet to a relying party's website. If the attacker attempts to present this captured data to a different website, the presentation will be rejected because it contains the original verifier's nonce and is signed with a pairwise key specific to the original verifier. Even replaying to the same verifier fails because the nonce has already been consumed and marked as used.
EUDI Wallet implementations also protect against cross-protocol replay attacks, where a presentation generated for one protocol (such as OpenID4VP) is replayed through a different protocol (such as direct API submission). The signed presentation includes the protocol context and intended audience in its payload, ensuring that presentations cannot be meaningfully repurposed across different verification protocols or verifier endpoints.