For authors

Prepare research that can be assessed.

A practical route through scope, manuscript files, scientific reporting and the declarations needed for a fair review.

AI and research integrity

Prepare an accurate AI-use declaration, verify references and explain the provenance of figures, data and code.

Read the practical guide

Manuscript Pre-check

Recommended before submission. Check sections, declarations, citations and references, then prepare a Word draft with this journal’s template.

Check your manuscript

Manuscript submissions

Submissions are currently closed while the editorial board is formed and the submission workflow is prepared. You may use the guidance and Manuscript Pre-check to prepare your work; preparation is not a journal submission. The opening date and authorised submission route will be announced here.

Author instructions and submission resources

Downloading a template or preparing evidence does not submit a manuscript. The formal submission route will be announced when the journal opens.

01

Check the fit

Identify the principal contribution and its fit with the journal’s scope.

02

Prepare the files

Separate the anonymised manuscript from author details.

03

Explain the evidence

Document methods, access conditions and limitations.

04

Complete declarations

Check authorship, ethics, support and related work.

Working Word resources

Use these editable files alongside the instructions below. They are preparation resources; they do not create an active submission.

WORD · DOCX

Manuscript template

Prepare the main manuscript and keep author identities in the separate title page.

Download manuscript template
WORD · DOCX

Title page

Record the author list, affiliations, contributions and relevant declarations separately.

Download title page
WORD · DOCX

Cover letter

Explain the contribution, journal fit and related work for editorial assessment.

Download cover letter

Preparing the submission files

Prepare an anonymised manuscript, a separate title page, a cover letter and any supporting files needed to assess the work. The title page should give the full journal title, manuscript title, article type, author names and affiliations, corresponding author details, contribution statement and relevant declarations. Keep identities and identifying acknowledgements out of the review manuscript. Anonymise document properties and explain any repository link that could reveal authorship. Ethics, data-access and methodological declarations needed to assess the science must remain in anonymised form in the review manuscript. Provide identifying approvals or other confidential evidence separately through the authorised editorial route.

Prepare an editable DOCX manuscript using the journal structure; a review PDF may accompany it. Authors using LaTeX or another specialist format should seek an agreed alternative route before submission and provide a compiled review PDF and source when requested. Supply original figure files and an accurate reference list for production. Initial submissions may use a consistent readable reference style. The final house style is specified during production. References to data and software should identify the version or release and a persistent identifier where available.

Manuscript structure and length

Use a specific title, an informative abstract and descriptive keywords. State the research question, its relationship to prior work, the methods and assumptions, the evidence, limitations and conclusions. For empirical work, distinguish results from interpretation. For theoretical, conceptual or qualitative work, use a structure that makes the reasoning and evidence traceable. Include all applicable declarations and cite supplementary material in the manuscript.

There is no fixed initial-submission word limit in this guide. Provide a word count and use the length needed for a complete, focused account. The editor may request shortening or restructuring. A later published article-type limit will apply to submissions received after its stated effective date; any exceptional change for a manuscript already under consideration must be communicated to its authors.

Submission and revision correspondence

A preparation file or saved evidence record does not constitute a journal submission. Formal submission requires the journal's announced submission route, author approval and completion of the required declarations. Authors should use the manuscript reference in subsequent correspondence and identify the version to which a query relates.

A revised submission should include an updated anonymised manuscript, any revised supporting files and a response addressing every editorial and reviewer point. Identify where changes were made and explain any reasoned disagreement. Keep the response professional and provide evidence for disputed technical issues. Do not silently substitute datasets, analyses or author lists; disclose consequential changes for editorial assessment.

Journal-specific standards

Scientific reporting

Use the guidance relevant to your design. A formal, conceptual or behavioural study should not add unrelated experiments merely to resemble another article type.

Threat models and security claims

Define the protected assets, relevant system boundaries, adversary capabilities and assumptions. Explain the access, knowledge and resources attributed to an attacker and what constitutes success or failure. State whether the work concerns prevention, detection, containment, privacy or recovery, and identify threats outside the analysis.

Align conclusions with those conditions. Protection against one attack or dataset does not establish general security. Discuss whether results depend on privileged information, unrealistic attacker limits, trusted components or an ability to update the system. Clearly distinguish properties demonstrated by analysis, experiments and observations from those proposed for future examination.

Authorisation and responsible disclosure

Describe the authority under which testing, data collection or access occurred, the relevant scope and any material restrictions. Where research involves third-party systems, evidence of authorisation may be requested for confidential editorial assessment. State how affected parties were engaged where applicable and explain any unresolved disclosure or safety concern.

Assess foreseeable harms from publishing vulnerability details, tools or datasets. Provide an appropriate disclosure chronology and explain any coordinated release, embargo or access limitation. The manuscript should support scientific assessment without unnecessarily enabling exploitation of a live weakness. The journal may seek further clarification or specialist advice and may defer a publication decision while material concerns remain unresolved.

Evaluation and deployment conditions

Justify datasets, comparison methods and metrics in relation to the threat model and intended use. Report class prevalence, labelling procedures, train and test separation, duplicate handling and any temporal or organisational partitioning. Explain contamination and leakage risks, including shared hosts, sessions, malware families or near-identical records where relevant.

Accuracy alone is rarely sufficient for a detection claim. Present relevant error rates, uncertainty and the consequences of false alarms or missed events. State resource costs, latency, capacity limits or analyst effort when deployment claims depend on them. Examine drift, important subgroups and failure conditions where the study supports such analysis; distinguish laboratory feasibility from demonstrated operational effectiveness.

Privacy and formal security evidence

Claims about privacy or cryptographic security should define the property, model and assumptions precisely. Identify what information is exposed, what parties are trusted and what adversarial behaviour is considered. Formal arguments should state their security notion, supporting assumptions and reasoning sufficiently for expert examination. Experimental resistance to selected attacks should not be presented as a general proof.

For differential privacy claims, specify the protected unit, neighbouring relation, privacy parameters and composition or accounting approach relevant to the release. Explain preprocessing and repeated releases that affect the claim. For anonymisation or synthetic-data studies, examine relevant disclosure risks and distinguish a reduced empirical attack rate from a formal privacy guarantee.

Sensitive resources and controlled release

Security resources may contain credentials, personal information, confidential configurations, malicious content or details of unresolved vulnerabilities. Explain collection rights, de-identification, handling and the access conditions appropriate to the material. Share only what can be released responsibly; a reproducibility objective does not justify exposing sensitive operational information.

When unrestricted release is inappropriate, describe a credible assessment route such as a sanitised sample, synthetic substitute, configuration description or controlled access arrangement. Explain which claims can be examined through that route and which remain limited. Identify resource versions, software dependencies, hardware, operating conditions and randomisation where material to the findings.

Human and organisational security research

Explain recruitment, participant characteristics, study setting, consent and ethics review or exemption as applicable. Security behaviour is sensitive to organisational authority, incentives and perceived consequences; describe measures protecting participants from employment, reputational or other foreseeable harm. Report survey or interview instruments and the analytical approach with appropriate confidentiality safeguards.

Where deception or incomplete disclosure is essential to the design, explain the ethical justification, approval, proportional safeguards and any debriefing arrangements. Do not assume that an organisational permission replaces consideration of individual participants. Distinguish self-reported intentions, simulated behaviour and observed operational practice, and avoid overstating what one setting establishes about other populations.

Incident evidence and anonymisation

Incident studies should identify the evidential basis, chronology and degree of access available to the researchers. Separate directly observed events, records provided by others and inferred explanations. Explain missing logs, competing hypotheses and uncertainty. A technically plausible reconstruction should not be reported as an established account without supporting evidence.

Remove unnecessary identifiers and details that could expose people, organisations or a continuing vulnerability. Explain how anonymisation affects interpretation and how material facts can be assessed confidentially where appropriate. Disclose employment, consultancy or other relationships with the affected organisation under the shared competing-interest policy.

AI security evaluation

Studies of attacks on or defences for AI systems should specify the model and data versions, the adversary's knowledge and access, and the evaluated attack conditions. Explain how attack success, clean performance, utility and relevant costs were measured. Compare defences under consistent conditions and discuss adaptive threats where they are relevant to the claim.

When AI supports security operations, evaluate the security task rather than treating generated explanations or fluent reports as reliable evidence. Describe output verification and consequential failure modes. An experiment on one model, benchmark or interaction pattern should not be extended to all deployments without a reasoned basis.

Declaration examples

Adapt these examples to what actually occurred. Keep science-relevant information in the anonymised manuscript and provide identifying evidence separately when needed.

Research authorisation

Testing and collection were conducted under [authorising body or agreement] within [scope]. The material restrictions were [restrictions]. Evidence supporting this statement can be provided to the editor through an appropriate confidential arrangement.

Vulnerability disclosure

The affected party was notified on [date] through [channel]. The current disclosure status is [status]. The manuscript and released materials omit or restrict [details] because [reason]. Any agreed release conditions are [conditions].

Restricted incident or security data

The study uses [records and period]. Public release is restricted because [specific reason]. The reviewed supporting materials and available verification route are [details]. These conditions limit assessment of [claims or components].

Threat model and evaluated environment

The evaluated adversary has [capabilities and access]. The protected assets and trusted components are [details]. Experiments use [system, data and resource versions]. Claims do not extend to [unevaluated conditions].

Common questions

Must every study introduce a new defence?

No. Measurement, theory, behavioural evidence, careful replication, negative results and incident analysis can advance cybersecurity when the contribution and supporting evidence are substantive.

Must exploit related material or sensitive traces be public?

No. Release should reflect authorisation, rights and foreseeable harm. Explain restrictions and provide an appropriate basis for assessment. Public availability and independent validation are separate matters.

Does using a public attack dataset establish ethical permission?

No. Explain the source, licence, collection context and relevant restrictions. Public availability does not remove questions about personal information, continuing vulnerabilities or inappropriate reuse.

Can an anonymised operational case be considered?

Yes, when the evidence can support meaningful assessment and the reasons for anonymisation are explained. Separate observation from inference and identify limitations caused by missing or restricted records.

Does successful peer review certify a system as secure?

No. A published study supports conclusions within its assumptions and evaluated conditions. The journal's editorial decision is not a security certification or a guarantee of safe deployment.