5G registration security gaps expose downgrade and replay risks

At a glance

What this is: NHI Mgmt Group's analysis, drawing on Bishop Fox's blog post: The 5G registration sequence still exposes an early unprotected window where downgrade, replay, and subscriber identifier risks can surface before NAS security is established.

Why it matters: It matters because 5G deployments that do not enforce strong negotiation, freshness, and concealed identifiers can turn protocol design into real-world exposure for mobile, OT, and private-network environments.

Context

5G registration is the point where a device first proves itself to the network and negotiates security for the session. The weakness is not the existence of cryptography later in the flow, but the fact that initial registration and authentication messages travel before NAS integrity and encryption are active, which creates a practical exposure window for attackers.

For IAM and NHI practitioners, this is a protocol-governance problem as much as a telecom one. The device, subscriber identity, and security-capability negotiation all behave like identity signals, and if the network accepts weak options or fails to conceal identifiers consistently, the result is exposure that looks similar to poor credential handling in other control planes.

Key questions

Q: What breaks when 5G registration allows weak security algorithms?

A: If a 5G core accepts null or weak algorithms during registration, the session can lose confidentiality and integrity before secure communication is established. That creates a downgrade path where attackers can observe or manipulate signaling, especially if the deployment does not enforce rejection of insecure options outside emergency use.

Q: Why do early 5G registration messages create replay and downgrade risk?

A: They are transmitted before NAS integrity and encryption are activated, so they rely on a vulnerable transition rather than full protection. That means attackers can capture, alter, or resend messages that the network has not yet bound to freshness or integrity. The risk is greatest where cores accept insecure negotiation or fail to validate context.

Q: What are the signs that a 5G core is mishandling registration security?

A: Warning signs include acceptance of EEA0 or EIA0 for ordinary traffic, repeated registration acceptance without clear freshness checks, and any SUPI or subscriber identifier appearing when SUCI should be used. Packet captures that show unprotected negotiation outcomes are especially important because they reveal the policy gap directly.

Q: How should operators test 5G registration security in private networks?

A: Test the full registration sequence, not only the authenticated end state. Focus on whether the core rejects weak algorithms, whether identifier concealment holds in live traffic, and whether replayed NAS messages are discarded. A good test plan combines packet capture, rogue infrastructure simulation, and policy validation against the early registration window.

Technical breakdown

Why the initial 5G registration window is exposed

The 5G registration flow begins with a Registration Request, then authentication exchange, and only later activates NAS security through the Security Mode Command and Complete. That sequencing means the first messages can be observed or influenced before integrity and confidentiality protections are in place. In security terms, the protocol uses a secure state transition, not a secure state from the first packet. If the core or access path tolerates weak handling in that transition, an attacker gains room to downgrade options, replay messages, or observe identity metadata.

Practical implication: Test and harden the unauthenticated phase explicitly, not just the secured phase that follows.

How algorithm negotiation becomes a downgrade path

During registration, the UE advertises supported encryption and integrity algorithms and the core decides what to accept. If the core allows null or legacy options such as EEA0 or EIA0 in normal conditions, the security posture drops to the weakest mutually accepted setting. A rogue gNB can exploit that negotiation by modifying signaling in transit, creating a bidding-down effect. The architectural issue is that negotiation is itself part of the attack surface, so policy enforcement has to happen at selection time, not after connection establishment.

Practical implication: Reject null and weak algorithms by policy before the session is admitted.

Why SUCI handling and replay resistance are both identity controls

5G replaces the permanent subscriber identifier with SUCI to conceal SUPI during transmission, but that only works if device, USIM, and core implementation align correctly. If concealment fails, the network leaks a stable identifier that can support tracking. Replay risk sits beside this problem because unauthenticated registration messages lack freshness guarantees until later in the flow. In practice, identifier protection and anti-replay controls are the same governance question: can the network prove a registration attempt is current and privacy-preserving before it accepts it?

Practical implication: Validate SUCI generation and replay resistance together during 5G testing.

Threat narrative

Attacker objective: The attacker aims to weaken or bypass registration security so they can expose subscriber identity, force insecure negotiation, or replay control-plane messages.

1. Entry occurs when an attacker intercepts or injects 5G registration traffic during the unauthenticated early exchange between UE and core.
2. Credential and identity handling are then weakened when the attacker exploits the cleartext window to observe or tamper with registration messages, including algorithm negotiation and subscriber identifiers.
3. Escalation follows when the attacker forces downgrade to null or weak encryption and integrity options, or replays a previously valid request that the core should have rejected.
4. Impact is exposure of subscriber identity, persistent tracking, or manipulated session establishment across the 5G control plane.

Breaches seen in the wild

• ShinyHunters FBI breach claim 2026: ShinyHunters claims it used a PeopleSoft flaw to breach the FBI and pivot into AWS GovCloud; the FBI has confirmed only an investigation.
• United Nations breach 2021: Sakura Samurai used exposed Git credentials to reach 100,000+ UNEP staff records, then reported the flaw through the UN disclosure programme.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 200+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

5G registration exposes a protocol-governance gap, not just a cryptographic one. The core issue is that identity and security negotiation begin before NAS protection is active, so the first trust decision happens in an exposed state. That means practitioners cannot treat later encryption as evidence that the earliest identity exchange was safe. The practical conclusion is that 5G security testing must target the transition into protection, not only the protected session.

Downgrade resilience is now a control-enforcement question. When a network accepts null or weak algorithms during registration, the attacker does not need to break cryptography. They only need to steer the session toward the weakest acceptable option. This is a familiar governance failure in a new protocol form: security capability exists, but policy does not consistently force it. Practitioners should treat algorithm negotiation as an enforced control boundary, not a compatibility feature.

SUCI handling is a subscriber-identity assurance problem with NHI parallels. A concealed identifier only helps if the concealment process is consistently implemented across device, USIM, and core. When it fails, the network exposes stable identity material that can be tracked, correlated, and reused. That is conceptually similar to poor machine-identity governance in other environments, where a supposedly protected identifier still leaks through process gaps. The practitioner takeaway is to govern identity concealment as a lifecycle control, not a one-time design choice.

Replay resistance belongs in the same assurance tier as authentication success. If a core cannot distinguish a fresh registration attempt from a captured one, the network accepts history as if it were current state. That breaks the trust model for mobile identity establishment and undermines any downstream access control that depends on registration integrity. In policy terms, freshness validation is not optional hardening. It is part of what makes the registration outcome trustworthy enough to use.

What this signals

Unauthenticated registration is the boundary that matters. Security teams often overrate the cryptography that arrives after session setup and underrate the trust decisions made before it. In 5G, that early boundary determines whether a core is merely capable of secure operation or actually enforcing it in real deployments.

5G identity governance needs the same discipline that IAM applies to privileged access. A concealed subscriber identifier is only useful when the concealment process, acceptance rules, and replay checks all behave consistently. The governance lesson is simple: identity protection has to be verified at issuance time, not assumed once the connection is live.

For practitioners

• Enforce strong algorithm baselines Configure 5G Core policy to reject null and weak encryption or integrity choices such as EEA0 and EIA0 in normal operation. Validate that admission control fails closed when a UE proposes insecure settings.
• Test the unprotected registration phase Build test cases for the Registration Request and Authentication Request/Response before NAS security activates. Verify that traffic cannot be modified, downgraded, or replayed without detection in that early window.
• Validate SUCI generation end to end Confirm that the UE, USIM, and home-network key provisioning consistently produce concealed identifiers before transmission. Watch for any SUPI leakage in packet captures or simulator output.

Key takeaways

• The central risk is the early 5G registration window, where security negotiation occurs before NAS protection is active.
• Bishop Fox’s testing shows that weak algorithm acceptance, replay acceptance, and identifier leakage can all emerge in that window if policy is loose.
• Operators need to validate rejection of insecure negotiation, SUCI handling, and freshness checks as part of 5G assurance.

Key terms

• NAS security: Non-Access Stratum security is the protection layer that secures signaling between the user equipment and the 5G core. In the registration flow, it is activated after the earliest messages, which is why those first exchanges require separate scrutiny and enforcement.
• SUCI: Subscription Concealed Identifier is the protected form of a subscriber identity used in 5G registration. It is designed to prevent plaintext exposure of the permanent subscriber identifier, but its privacy value depends on correct device implementation and key provisioning.
• Downgrade attack: A downgrade attack forces a connection to use a weaker protocol or cipher than both sides could otherwise support. In practice, it turns legacy compatibility into a security weakness by increasing the chance that known flaws or weaker cryptography can be exploited during negotiation.
• Replay Attack: An attack where a previously valid signed message is captured and sent again later to trigger the same trusted action. Timestamping, nonces, and context binding are the main controls that reduce replay risk in HMAC-based systems.

Deepen your knowledge

NHI Mgmt Group covers identity security, NHI governance, and agentic AI through independent research, practitioner guides, and the NHI Foundation Level course. It is designed for practitioners who need to connect identity governance to real operational control decisions.