Chris J. Mitchell

Chris J Mitchell

Three closely-related polynomial-based group key pre-distribution schemes have recently been proposed, aimed specifically at wireless sensor networks. The schemes enable any subset of a predefined set of sensor nodes to establish a shared secret key without any communications overhead. It is claimed that these schemes are both secure and lightweight, i.e. making them particularly appropriate for network scenarios where nodes have limited computational and storage capabilities. Further papers have built on these schemes, e.g. to propose secure routing protocols for wireless sensor networks. Unfortunately, as we show in this paper, all three schemes are completely insecure; whilst the details of their operation varies, they share common weaknesses. In every case we show that an attacker equipped with the information built into at most two sensor nodes can compute group keys for all possible groups of which the attacked nodes are not a member, which breaks a fundamental design objective. The attacks can also be achieved by an attacker armed with the information from a single node together with a single group key to which this sensor node is not entitled. Repairing the schemes appears difficult, if not impossible. The existence of major flaws is not surprising given the complete absence of any rigorous proofs of security for the proposed schemes. A further recent paper proposes a group membership authentication and key establishment scheme based on one of the three key pre-distribution schemes analysed here; as we demonstrate, this scheme is also insecure, as the attack we describe on the corresponding pre-distribution scheme enables the authentication process to be compromised.

Swapnil Bhosale, Cosmin Frateanu, Camilla Clark et al.

Deploying accurate event detection on resource-constrained devices is challenged by the trade-off between performance and computational cost. While Early-Exit (EE) networks offer a solution through adaptive computation, they often fail to enforce a coherent hierarchical structure, limiting the reliability of their early predictions. To address this, we propose Hyperbolic Early-Exit networks (HypEE), a novel framework that learns EE representations in the hyperbolic space. Our core contribution is a hierarchical training objective with a novel entailment loss, which enforces a partial-ordering constraint to ensure that deeper network layers geometrically refine the representations of shallower ones. Experiments on multiple audio event detection tasks and backbone architectures show that HypEE significantly outperforms standard Euclidean EE baselines, especially at the earliest, most computationally-critical exits. The learned geometry also provides a principled measure of uncertainty, enabling a novel triggering mechanism that makes the overall system both more efficient and more accurate than a conventional EE and standard backbone models without early-exits.

Haitham S. Al-Sinani, Chris J. Mitchell

Traditional ethical hacking relies on skilled professionals and time-intensive command management, which limits its scalability and efficiency. To address these challenges, we introduce PenTest++, an AI-augmented system that integrates automation with generative AI (GenAI) to optimise ethical hacking workflows. Developed in a controlled virtual environment, PenTest++ streamlines critical penetration testing tasks, including reconnaissance, scanning, enumeration, exploitation, and documentation, while maintaining a modular and adaptable design. The system balances automation with human oversight, ensuring informed decision-making at key stages, and offers significant benefits such as enhanced efficiency, scalability, and adaptability. However, it also raises ethical considerations, including privacy concerns and the risks of AI-generated inaccuracies (hallucinations). This research underscores the potential of AI-driven systems like PenTest++ to complement human expertise in cybersecurity by automating routine tasks, enabling professionals to focus on strategic decision-making. By incorporating robust ethical safeguards and promoting ongoing refinement, PenTest++ demonstrates how AI can be responsibly harnessed to address operational and ethical challenges in the evolving cybersecurity landscape.

Haitham S. Al-Sinani, Chris J. Mitchell

This study explores the application of generative AI (GenAI) within manual exploitation and privilege escalation tasks in Linux-based penetration testing environments, two areas critical to comprehensive cybersecurity assessments. Building on previous research into the role of GenAI in the ethical hacking lifecycle, this paper presents a hands-on experimental analysis conducted in a controlled virtual setup to evaluate the utility of GenAI in supporting these crucial, often manual, tasks. Our findings demonstrate that GenAI can streamline processes, such as identifying potential attack vectors and parsing complex outputs for sensitive data during privilege escalation. The study also identifies key benefits and challenges associated with GenAI, including enhanced efficiency and scalability, alongside ethical concerns related to data privacy, unintended discovery of vulnerabilities, and potential for misuse. This work contributes to the growing field of AI-assisted cybersecurity by emphasising the importance of human-AI collaboration, especially in contexts requiring careful decision-making, rather than the complete replacement of human input.

Haitham S. Al-Sinani, Chris J. Mitchell

This paper investigates the presence and impact of questionable, AI-generated academic papers on widely used preprint repositories, with a focus on their role in citation manipulation. Motivated by suspicious patterns observed in publications related to our ongoing research on GenAI-enhanced cybersecurity, we identify clusters of questionable papers and profiles. These papers frequently exhibit minimal technical content, repetitive structure, unverifiable authorship, and mutually reinforcing citation patterns among a recurring set of authors. To assess the feasibility and implications of such practices, we conduct a controlled experiment: generating a fake paper using GenAI, embedding citations to suspected questionable publications, and uploading it to one such repository (ResearchGate). Our findings demonstrate that such papers can bypass platform checks, remain publicly accessible, and contribute to inflating citation metrics like the H-index and i10-index. We present a detailed analysis of the mechanisms involved, highlight systemic weaknesses in content moderation, and offer recommendations for improving platform accountability and preserving academic integrity in the age of GenAI.

Gaëtan Pradel, Chris Mitchell

Biometric matching involves storing and processing sensitive user information. Maintaining the privacy of this data is thus a major challenge, and homomorphic encryption offers a possible solution. We propose a privacy-preserving biometrics-based authentication protocol based on fully homomorphic encryption, where the biometric sample for a user is gathered by a local device but matched against a biometric template by a remote server operating solely on encrypted data. The design ensures that 1) the user's sensitive biometric data remains private, and 2) the user and client device are securely authenticated to the server. A proof-of-concept implementation building on the TFHE library is also presented, which includes the underlying basic operations needed to execute the biometric matching. Performance results from the implementation show how complex it is to make FHE practical in this context, but it appears that, with implementation optimisations and improvements, the protocol could be used for real-world applications.

Chris J Mitchell

Two recently published papers propose some very simple key distribution schemes designed to enable two or more parties to establish a shared secret key with the aid of a third party. Unfortunately, as we show, most of the schemes are inherently insecure and all are incompletely specified - moreover, claims that the schemes are inherently lightweight are shown to be highly misleading. We also briefly critique a somewhat related very recent paper by the same authors that uses similar techniques to achieve what are claimed to be secure multiparty computations.

Chris J Mitchell

Serious weaknesses in two very closely related group authentication and group key establishment schemes are described. Simple attacks against the group key establishment part of the schemes are described, which strongly suggest that the schemes should not be used.

Chris J Mitchell

A paper presented at the ICICS 2019 conference describes what is claimed to be a `provably secure group authentication [protocol] in the asynchronous communication model'. We show here that this is far from being the case, as the protocol is subject to serious attacks. To try to explain this troubling case, an earlier (2013) scheme on which the ICICS 2019 protocol is based was also examined and found to possess even more severe flaws - this latter scheme was previously known to be subject to attack, but not in quite as fundamental a way as is shown here. Examination of the security theorems provided in both the 2013 and 2019 papers reveals that in neither case are they exactly what they seem to be at first sight; the issues raised by this are also briefly discussed.

Chris J Mitchell

Two recent papers describe almost exactly the same group key establishment protocol for wireless sensor networks. Quite part from the duplication issue, we show that both protocols are insecure and should not be used - a member of a group can successfully impersonate the key generation centre and persuade any other group member to accept the wrong key value. This breaks the stated objectives of the schemes.

Chris J Mitchell

There has been much recent discussion of the criticality of the 5G infrastructure, and whether certain vendors should be able to supply 5G equipment. The key issue appears to be about trust, namely to what degree the security and reliability properties of 5G equipment and systems need to be trusted, and by whom, and how the necessary level of trust might be obtained. In this paper, by considering existing examples such as the Internet, the possible need for trust is examined in a systematic way, and possible routes to gaining trust are described. The issues that arise when a security and/or reliability failure actually occurs are also discussed. The paper concludes with a discussion of possible future ways of enabling all parties to gain the assurances they need in a cost-effective and harmonised way.

Chris J Mitchell

Three closely-related polynomial-based group key pre-distribution schemes have recently been proposed, aimed specifically at wireless sensor networks. The schemes enable any subset of a predefined set of sensor nodes to establish a shared secret key without any communications overhead. It is claimed that these schemes are both secure and lightweight, i.e. making them particularly appropriate for network scenarios where nodes have limited computational and storage capabilities. Further papers have built on these schemes, e.g. to propose secure routing protocols for wireless sensor networks. Unfortunately, as we show in this paper, all three schemes are completely insecure; whilst the details of their operation varies, they share common weaknesses. In every case we show that an attacker equipped with the information built into at most two sensor nodes can compute group keys for all possible groups of which the attacked nodes are not a member, which breaks a fundamental design objective. The attacks can also be achieved by an attacker armed with the information from a single node together with a single group key to which this sensor node is not entitled. Repairing the schemes appears difficult, if not impossible. The existence of major flaws is not surprising given the complete absence of any rigorous proofs of security for the proposed schemes. A further recent paper proposes a group membership authentication and key establishment scheme based on one of the three key pre-distribution schemes analysed here; as we demonstrate, this scheme is also insecure, as the attack we describe on the corresponding pre-distribution scheme enables the authentication process to be compromised.

Chris J Mitchell

A recently proposed group key distribution scheme known as UMKESS, based on secret sharing, is shown to be insecure. Not only is it insecure, but it does not always work, and the rationale for its design is unsound. UMKESS is the latest in a long line of flawed group key distribution schemes based on secret sharing techniques.

Chris J Mitchell

This paper provides a detailed analysis of the impact of quantum computing on the security of 5G mobile telecommunications. This involves considering how cryptography is used in 5G, and how the security of the system would be affected by the advent of quantum computing. This leads naturally to the specification of a series of simple, phased, recommended changes intended to ensure that the security of 5G (as well as 3G and 4G) is not badly damaged if and when large scale quantum computing becomes a practical reality. By exploiting backwards-compatibility features of the 5G security system design, we are able to propose a novel multi-phase approach to upgrading security that allows for a simple and smooth migration to a post-quantum-secure system.

Chris J Mitchell

A recently proposed authenticated key agreement protocol is shown to be insecure. In particular, one of the two parties is not authenticated, allowing an active man in the middle opponent to replay old messages. The protocol is essentially an authenticated Diffie-Hellman key agreement scheme, and the lack of authentication allows an attacker to replay old messages and have them accepted. Moreover, if the ephemeral key used to compute a protocol message is ever compromised, then the key established using the replayed message will also be compromised. Fixing the problem is simple - there are many provably secure and standardised protocols which are just as efficient as the flawed scheme.

Wanpeng Li, Chris J Mitchell, Thomas Chen

Millions of users routinely use Google to log in to websites supporting OAuth 2.0 or OpenID Connect; the security of OAuth 2.0 and OpenID Connect is therefore of critical importance. As revealed in previous studies, in practice RPs often implement OAuth 2.0 incorrectly, and so many real-world OAuth 2.0 and OpenID Connect systems are vulnerable to attack. However, users of such flawed systems are typically unaware of these issues, and so are at risk of attacks which could result in unauthorised access to the victim user's account at an RP. In order to address this threat, we have developed OAuthGuard, an OAuth 2.0 and OpenID Connect vulnerability scanner and protector, that works with RPs using Google OAuth 2.0 and OpenID Connect services. It protects user security and privacy even when RPs do not implement OAuth 2.0 or OpenID Connect correctly. We used OAuthGuard to survey the 1000 top-ranked websites supporting Google sign-in for the possible presence of five OAuth 2.0 or OpenID Connect security and privacy vulnerabilities, of which one has not previously been described in the literature. Of the 137 sites in our study that employ Google Sign-in, 69 were found to suffer from at least one serious vulnerability. OAuthGuard was able to protect user security and privacy for 56 of these 69 RPs, and for the other 13 was able to warn users that they were using an insecure implementation.

Chris J Mitchell

A significant security vulnerability in a recently published group key establishment protocol is described. This vulnerability allows a malicious insider to fraudulently establish a group key with an innocent victim, with the key chosen by the attacker. This shortcoming is sufficiently serious that the protocol should not be used.

Chris J Mitchell

Major shortcomings in a recently published group key establishment protocol are described. These shortcomings are sufficiently serious that the protocol should not be used.

Wanpeng Li, Chris J Mitchell, Thomas Chen

Many millions of users routinely use their Google, Facebook and Microsoft accounts to log in to websites supporting OAuth 2.0 and/or OpenID Connect-based single sign on. The security of OAuth 2.0 and OpenID Connect is therefore of critical importance, and it has been widely examined both in theory and in practice. Unfortunately, as these studies have shown, real-world implementations of both schemes are often vulnerable to attack, and in particular to cross-site request forgery (CSRF) attacks. In this paper we propose a new technique which can be used to mitigate CSRF attacks against both OAuth 2.0 and OpenID Connect.

Fatma Al Maqbali, Chris J Mitchell

Web password recovery, enabling a user who forgets their password to re-establish a shared secret with a website, is very widely implemented. However, use of such a fall-back system brings with it additional vulnerabilities to user authentication. This paper provides a framework within which such systems can be analysed systematically, and uses this to help gain a better understanding of how such systems are best implemented. To this end, a model for web password recovery is given, and existing techniques are documented and analysed within the context of this model. This leads naturally to a set of recommendations governing how such systems should be implemented to maximise security. A range of issues for further research are also highlighted.

Fatma Al Maqbali, Chris J Mitchell

Text password has long been the dominant user authentication technique and is used by large numbers of Internet services. If they follow recommended practice, users are faced with the almost insuperable problem of generating and managing a large number of site-unique and strong (i.e. non-guessable) passwords. One way of addressing this problem is through the use of a password generator, i.e. a client-side scheme which generates (and regenerates) site-specific strong passwords on demand, with the minimum of user input. This paper provides a detailed specification and analysis of AutoPass, a password generator scheme previously outlined as part of a general analysis of such schemes. AutoPass has been designed to address issues identified in previously proposed password generators, and incorporates novel techniques to address these issues. Unlike almost all previously proposed schemes, AutoPass enables the generation of passwords that meet important real-world requirements, including forced password changes, use of pre-specified passwords, and generation of passwords meeting site-specific requirements.

Fatma AL Maqbali, Chris J Mitchell

This paper considers password generators, i.e. systems designed to generate site-specific passwords on demand. Such systems are an alternative to password managers. Over the last 15 years a range of password generator systems have been described. This paper proposes the first general model for such systems, and critically examines options for instantiating this model; options considered include all those previously proposed as part of existing schemes as well as certain novel possibilities. The model enables a more objective and high-level assessment of the design of such systems; it has also been used to sketch a possible new scheme, AutoPass, intended to incorporate the best features of the prior art whilst also addressing many of the most serious shortcomings of existing systems through the inclusion of novel features.

Mohammed Shafiul Alam Khan, Chris J Mitchell

As has been widely discussed, the GSM mobile telephony system only offers unilateral authentication of the mobile phone to the network; this limitation permits a range of attacks. While adding support for mutual authentication would be highly beneficial, changing the way GSM serving networks operate is not practical. This paper proposes a novel modification to the relationship between a Subscriber Identity Module (SIM) and its home network which allows mutual authentication without changing any of the existing mobile infrastructure, including the phones; the only necessary changes are to the authentication centres and the SIMs. This enhancement, which could be deployed piecemeal in a completely transparent way, not only addresses a number of serious vulnerabilities in GSM but is also the first proposal for enhancing GSM authentication that possesses such transparency properties.

Chris J Mitchell

This paper reconsiders the security offered by 2-key triple DES, an encryption technique that remains widely used despite recently being de-standardised by NIST. A generalisation of the 1990 van Oorschot-Wiener attack is described, constituting the first advance in cryptanalysis of 2-key triple DES since 1990. We give further attack enhancements that together imply that the widely used estimate that 2-key triple DES provides 80 bits of security can no longer be regarded as conservative; the widely stated assertion that the scheme is secure as long as the key is changed regularly is also challenged. The main conclusion is that, whilst not completely broken, the margin of safety for 2-key triple DES is slim, and efforts to replace it, at least with its 3-key variant, should be pursued with some urgency.

Wanpeng Li, Chris J Mitchell

Many millions of users routinely use their Google accounts to log in to relying party (RP) websites supporting the Google OpenID Connect service. OpenID Connect, a newly standardised single-sign-on protocol, builds an identity layer on top of the OAuth 2.0 protocol, which has itself been widely adopted to support identity management services. It adds identity management functionality to the OAuth 2.0 system and allows an RP to obtain assurances regarding the authenticity of an end user. A number of authors have analysed the security of the OAuth 2.0 protocol, but whether OpenID Connect is secure in practice remains an open question. We report on a large-scale practical study of Google's implementation of OpenID Connect, involving forensic examination of 103 RP websites which support its use for sign-in. Our study reveals serious vulnerabilities of a number of types, all of which allow an attacker to log in to an RP website as a victim user. Further examination suggests that these vulnerabilities are caused by a combination of Google's design of its OpenID Connect service and RP developers making design decisions which sacrifice security for simplicity of implementation. We also give practical recommendations for both RPs and OPs to help improve the security of real world OpenID Connect systems.

Mohammed Shafiul Alam Khan, Chris J Mitchell

Although the security properties of 3G and 4G mobile networks have significantly improved by comparison with 2G (GSM), significant shortcomings remain with respect to user privacy. A number of possible modifications to 2G, 3G and 4G protocols have been proposed designed to provide greater user privacy; however, they all require significant modifications to existing deployed infrastructures, which are almost certainly impractical to achieve in practice. In this article we propose an approach which does not require any changes to the existing deployed network infrastructures or mobile devices, but offers improved user identity protection over the air interface. The proposed scheme makes use of multiple IMSIs for an individual USIM to offer a degree of pseudonymity for a user. The only changes required are to the operation of the authentication centre in the home network and to the USIM, and the scheme could be deployed immediately since it is completely transparent to the existing mobile telephony infrastructure. We present two different approaches to the use and management of multiple IMSIs.