Securing the open-source ecosystem: AI

Modern software supply chains face increasing risks from vulnerable and anomalous dependencies, necessitating automated and interpretable detection methods integrated within Continuous Integration and Continuous Deployment (CI/CD) workflows. Open-source software (OSS) ecosystems are increasingly targeted by sophisticated attacks—including AI model poisoning, dependency confusion, contributor hijacking, and transitive dependency anomalies—that often evade traditional CVE-centric tools such as Snyk and Dependabot. These reactive solutions are primarily designed for known vulnerabilities and may exhibit high false-positive rates when applied to behavioral anomalies, while also lacking interpretable and actionable guidance for developers. This limitation creates a critical barrier to reliable and secure software development in modern DevSecOps environments. To address this gap, this paper introduces XAI-SCS, an intelligent, explainable decision-support framework for OSS supply chain security that performs metadata-driven behavioral risk assessment to proactively identify anomalous package activity. A hybrid CNN–LSTM–Autoencoder architecture achieves strong anomaly-detection performance on the held-out test set (F1 = 0.91), with low repeated-run variability (mean F1 = 0.90 ± 0.02 across five runs) on a federated dataset of 12,000 npm/PyPI package versions, primarily trained using CVE-linked labels and GAN-simulated poisoning samples. A separate held-out evaluation set of 240 manually curated real-world non-CVE OSS supply-chain incidents was used exclusively for external generalization assessment. Accordingly, the non-CVE results are interpreted as preliminary evidence of transferability to selected metadata-visible behavioral anomalies rather than comprehensive real-world non-CVE threat coverage. To support transparent developer decision-making, SHAP-based explanations and counterfactual remediation suggestions produced high perceived clarity (4.3 ± 0.4/5) and actionability (4.1 ± 0.5/5) and self-reported remediation intent (96%, n = 82). However, these findings reflect developer perceptions in a controlled study and do not yet demonstrate measured improvements in real-world remediation time, fix completion, or vulnerability resolution within operational DevSecOps teams. Optional high-assurance deployment extensions, including Federated Continual Learning, zero-knowledge attestation via zk-SNARKs, and post-quantum secure enclaves, are presented as exploratory architectural pathways for future privacy-preserving, verifiable, and confidential deployment, rather than as part of the core empirical evaluation. Embedded in controlled, CI/CD-representative workflows with 0.8-second average latency, XAI-SCS remains competitive with CVE-centric tools on known-vulnerability cases while providing preliminary evidence of complementary metadata-level behavioral risk screening in GAN-simulated and limited curated non-CVE evaluation subsets. The comparison is therefore interpreted as evidence that XAI-SCS can complement CVE-centric SCA tools by adding metadata-level behavioral screening and developer-facing explanations, rather than replacing or universally outperforming tools such as Snyk and Dependabot. This work’s primary contribution is a metadata-driven, CI/CD-integrated intelligent decision-support framework that delivers high-accuracy behavioral anomaly detection (F1 = 0.91) and transparent, developer-facing explanations (clarity 4.3/5, remediation intent 96%) for OSS supply chain risk assessment. Optional high-assurance deployment extensions (federated learning, cryptographic attestation, secure enclaves) are discussed separately and do not form part of the core empirical evaluation.

Similar content being viewed by others

AI-driven cybersecurity framework for anomaly detection in power systems

A change-aware, AI-assisted framework for dynamic third-party cybersecurity risk assessment

Explainable dual LSTM-autoencoders with exogenous features for anomaly detection and supply chain forecasting

This research is supported by a grant (No. CRPG-25-3096) under the Cybersecurity Research and Innovation Pioneers Initiative, provided by the National Cybersecurity Authority (NCA) in the Kingdom of Saudi Arabia.

Authors and Affiliations

Department of Computer Science, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, Saudi Arabia Ahmed Alzahrani

Department of Computer Science, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, Saudi Arabia

Gomal Research Institute of Computing (GRIC), Faculty of Computing, Gomal University, D. I. Khan, Khyber Pakhtunkhwa, Pakistan Muhammad Zubair Asghar

Gomal Research Institute of Computing (GRIC), Faculty of Computing, Gomal University, D. I. Khan, Khyber Pakhtunkhwa, Pakistan

Muhammad Zubair Asghar

Ahmed Alzahrani View author publications author on: PubMed Google Scholar

author on: PubMed Google Scholar

Muhammad Zubair Asghar View author publications author on: PubMed Google Scholar

author on: PubMed Google Scholar

Corresponding authors

Correspondence to Ahmed Alzahrani or Muhammad Zubair Asghar .

The authors declare no competing interests.

Declaration of AI tool usage

During the preparation of this manuscript, the author(s) used the AI-based tools Grammarly, ChatGPT, Grok, and Gemini exclusively for correcting grammatical errors and enhancing the professional quality of the language. After utilizing these tools, the author(s) thoroughly reviewed and edited the content to ensure accuracy and appropriateness, and take(s) full responsibility for the final content of the publication.

This study involved two distinct methodological components. First, the software supply-chain dataset was constructed exclusively from publicly accessible sources, including npm, PyPI, GitHub APIs, and the National Vulnerability Database (NVD). These sources contain package-level metadata, dependency structures, contributor activity indicators, and vulnerability records, and do not include personally identifiable information, private user data, or restricted-access content. Therefore, no separate Institutional Review Board (IRB) approval or individual informed consent was required for the public dataset component. Second, the developer survey conducted for the explainability evaluation involved voluntary human participants and was formally approved under Ethics Approval No. KAU-CS-2025-03. Participation was voluntary and anonymous, informed consent was obtained prior to participation, and no personally identifiable information was collected. Additional details regarding the survey procedure, consent process, and anonymized responses are provided in Appendix A.

Additional information

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1

Appendix A: Developer survey, consent procedure, and ethics approval (KAU-CS-2025-03)

Introduction and consent

This survey evaluates perceptions of XAI-SCS, an explainable-AI framework for open-source supply-chain security.

Participation is voluntary and anonymous; the estimated time is 8–10 min. By proceeding, you consent to participate under the approved protocol.

☐ I consent to participate Date: ____________________.

Background information (demographics)

Q1. Years of experience

☐ 10

Q2. Primary package manager

☐ npm ☐ PyPI ☐ Both ☐ Other

Q3. Use CI/CD pipelines regularly ?

☐ Developer ☐ Security Engineer ☐ DevOps ☐ Maintainer ☐ Researcher

Explanation clarity (5-point likert)

Scale: 1 = Strongly Disagree 2 = Disagree 3 = Neutral 4 = Agree 5 = Strongly Agree.

SHAP explanation clearly shows which dependency features drive the risk score.

Counterfactual (“What-if”) message is understandable and relevant.

Explanation format integrates naturally within CI/CD logs.

Technical detail level is appropriate for my expertise.

Visualization (bar plot / heatmap) helps interpret the alert cause.

Actionability and practical utility

Counterfactual guidance specifies clear remediation steps.

I would act on most XAI-SCS alerts owing to clarity of recommendations.

XAI-SCS appears likely to help shorten remediation time compared with conventional alert-only tools.

Recommendations align with organizational CI/CD policy.

Alerts can be integrated easily into my workflow.

Trust, transparency and confidence

I trust the risk score because the explanation clearly shows the main contributing features.

The recommendation is understandable enough to support review or remediation.

The CI/CD report format increases my confidence in the alert.

Explanations make the model’s reasoning transparent and fair.

Overall, XAI-SCS is trustworthy for OSS risk assessment.

T3: The CI/CD report format increases my confidence in the alert.

Intent and future use

How often would you act on XAI-SCS alerts?

☐ Always ☐ Often ☐ Sometimes ☐ Rarely ☐ Never

Would you recommend XAI-SCS to others?

Preferred alert delivery channel

☐ CI/CD log ☐ Email ☐ Slack ☐ Dashboard widget

Open-ended feedback (optional)

1. What was most useful the explanations? __________________________________.

What was most useful the explanations? __________________________________.

2. What improvements would make alerts more actionable? __________________________.

What improvements would make alerts more actionable? __________________________.

3. Do you have any privacy or bias concerns regarding AI-based alerts? ______________.

Do you have any privacy or bias concerns regarding AI-based alerts? ______________.

Arabic version (النسخة العربية)

(Right-to-Left layout recommended in Word) .

القسم الأول – المعلومات الأساسية .

س1. سنوات الخبرة … ☐ أقل من سنتين ☐ ٢–٥ ☐ ٦–١٠ ☐ أكثر من ١٠.

س2. مدير الحزم … ☐ npm ☐ PyPI ☐ كلاهما ☐ آخر.

س3. هل تستخدم خطوط CI/CD بشكل منتظم؟ ☐ نعم ☐ لا.

س4. الدور الوظيفي ☐ مطور ☐ مهندس أمن ☐ DevOps ☐ مسؤول صيانة ☐ باحث.

القسم الثاني – وضوح الشرح .

C1 تُظهر SHAP … C5 التصور البصري يساعد في الفهم.

القسم الثالث – القابلية للتنفيذ .

A1 الشرح يقدم خطوات … A5 سهولة الدمج في سير العمل.

القسم الرابع – الثقة والشفافية .

T1 التحقق باستخدام zk-SNARK … T5 أجد XAI-SCS موثوقًا لتقييم المخاطر.

القسم الخامس – النوايا والاستخدام المستقبلي .

I1 كم مرة … I3 ما الطريقة المفضلة لتلقي التنبيهات ؟.

القسم السادس – ملاحظات إضافية (اختياري) .

All responses were collected anonymously between March 10–25 2025. Anonymized data and analysis scripts are available .

Appendix B. Optional High-Assurance Deployment Extensions

This appendix briefly summarizes optional deployment extensions that may be explored in future work. These extensions are not part of the primary empirical evaluation of XAI-SCS.

Privacy-preserving adaptation

Future deployments may use privacy-preserving model adaptation to update risk-scoring models across multiple registries without directly sharing raw package metadata. In this manuscript, this idea is treated only as an exploratory deployment direction.

Provenance verification

Future deployments may attach cryptographic provenance checks to SBOMs or risk-score outputs to support tamper-evident reporting. The present study does not evaluate cryptographic soundness or production-grade attestation.

Confidential execution

Future deployments may use secure-execution environments to protect inference in untrusted CI/CD runners. The present study does not provide a full enclave-security evaluation or post-quantum security proof.

These optional extensions are included only to indicate possible future high-assurance deployment pathways. They are subordinate to the core XAI-SCS contribution, which remains the metadata-driven anomaly-detection, explanation, and CI/CD decision-support framework evaluated under RQ1–RQ3.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit .

Reprints and permissions

Alzahrani, A., Asghar, M.Z. Securing the open-source ecosystem: AI-enhanced and explainable supply chain security for reliable software development. Sci Rep (2026).

Received : 10 January 2026

Received : 10 January 2026

Accepted : 29 August 2026

Accepted : 29 August 2026

Published : 03 October 2026

Published : 03 October 2026

DOI :

DOI :

Intelligent decision-support system

Explainable artificial intelligence

Software supply chain security

Metadata-based anomaly detection