cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Mardalius
Contact Email
mardalius18@gmail.com
Phone
+6282284039993
Journal Mail Official
mardalius18@gmail.com
Editorial Address
Jl. Rambe, Perumahan Pelita Bangsa, Blok A, No.15, Kab. Asahan, Sumatera Utara
Location
Kab. asahan,
Sumatera utara
INDONESIA
Journal Of Computer Science And Technology
Published by PT. Padang Tekno Corp
ISSN : 29855772     EISSN : 29854318     DOI : https://doi.org/10.59435/jocstec
Core Subject : Science,
Computer Science & IT
Journal of Computer Science And Technology (JOCSTEC) is a scientific journal that publishes research results and thoughts in the field of computers and information technology. JOCSTEC focuses on publishing research results that contribute to understanding theory and applications in the field of computers and information technology, including computer systems, computer networks, data processing, information security, and other information technologies.
Arjuna Subject : Ilmu Komputer (semua kategori) - Ilmu Komputer (Lain-Lain)
Articles 83 Documents
 5   6   7   8   9 
Cross-Lingual Indirect Prompt Injection Across Retrieval, Reranking, And Generation In Multilingual RAG Fauzi Bondan Prihananto; Erlangga Bayu Yudho Prakoso; Aprilisa Arum Sari; Tomy Anugrah Islami
Journal of Computer Science and Technology (JOCSTEC) Vol 4 No 3 (2026): JOCSTEC - September
Publisher : PT. Padang Tekno Corp

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59435/jocstec.v4i3.809

Abstract

External evidence can make retrieval-augmented generation (RAG) more informative, yet retrieved passages also provide a path for adversarial instructions to enter the model context. We examine that path in an English-Indonesian RAG system and track cross-lingual indirect prompt injection separately at retrieval, reranking, and generation. The experiment starts from 75 semantic items and evaluates every item in eight query/body/payload language combinations, for 600 paired trials. Qwen3-Embedding-0.6B and BGE-M3 produce top-20 candidate sets, BGE-reranker-v2-m3 reduces each set to five documents, and Qwen3-0.6B answers from the resulting context with either a standard prompt or an explicit trust-boundary prompt. Statistical uncertainty is estimated by resampling semantic items, and paired binary outcomes are modeled with generalized estimating equations. Poison documents reached the top 20 in 80.50% of Qwen trials and 37.67% of BGE-M3 trials (odds ratio 6.94, 95% CI 4.53-10.63). Top-five exposure was 18.50% and 16.17%, respectively. Standard end-to-end attack success was 6.33% for Qwen and 6.00% for BGE-M3; boundary-aware prompting lowered both rates to 1.33%, with no canary false positives. The results indicate that multilingual RAG security depends on several linked stages rather than generation alone.
Design And Development Of A Web-Based Examination Application Using Django Framework At SMA Negeri 7 Sarolangun Afif Muhammad; Thomson Mary; Ami Anggraini Samudra
Journal of Computer Science and Technology (JOCSTEC) Vol 4 No 3 (2026): JOCSTEC - September
Publisher : PT. Padang Tekno Corp

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59435/jocstec.v4i3.833

Abstract

The implementation of examinations at SMA Negeri 7 Sarolangun previously encountered operational challenges due to an unintegrated workflow, where exam authoring was conducted via Google Forms and exam delivery relied on an Android-based third-party application paired with a custom keyboard. This workflow caused severe device incompatibility, system crashes, and manual score recapitulation delays lasting up to one week. This research aims to design and develop an integrated web-based examination application utilizing the Django framework under the Rapid Application Development (RAD) model to support Midterm (PTS) and Final (PAS) semester exams. The application features master data management, manual question authoring and bulk imports via Excel, token-protected scheduling, asynchronous AJAX auto-saving on computer-based test (CBT) sheets, live proctoring controls, auto-grading, score visibility authorization (Hide Score), and dynamic class-filtered report exports in Excel and PDF formats. System evaluation was conducted through Alpha testing (White Box and Black Box testing) and Beta testing (ISO/IEC 25010:2011 quality evaluation by system experts and End-User Computing Satisfaction/EUCS by end-users). Alpha testing verified that all 29 independent logic paths in White Box testing executed without branching errors and Black Box testing achieved 100% validity across 13 functional modules. Beta testing with 2 system experts yielded an average score of 98.25% (Very Good), and end-user testing involving 27 respondents achieved an average satisfaction score of 84.75% (Very Good). Thus, the developed system is highly feasible, reliable, and effective for school examination operations.
Design And Development Of A Laravel 13 Based E-Commerce Referral System For Seller Monitoring And Accountability Tegar Puji Bachtiar; Primaadi Airlangga
Journal of Computer Science and Technology (JOCSTEC) Vol 4 No 3 (2026): JOCSTEC - September
Publisher : PT. Padang Tekno Corp

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59435/jocstec.v4i3.854

Abstract

The rapid growth of electronic transactions necessitates systems capable of monitoring seller quality and reputation across marketplace platforms. This study aims to design and implement a seller referral system on a Laravel 13 based e-commerce platform integrated with performance monitoring and seller accountability mechanisms. Built using the Prototype model encompassing communication, planning, modeling, prototyping, and deployment phases the system utilizes MySQL for database management. Registered sellers can distribute unique referral codes to prospective sellers, with the system automatically recording these relationships to enable structured tracking of activities, product counts, store ratings, and access status of referred sellers. Functional testing conducted via a black-box approach demonstrated that the authentication, referral registration, dashboard, monitoring, order management, and store settings modules operated strictly according to specifications, achieving a 100% test case pass rate. The primary contribution of this research lies in shifting the referral paradigm from simple user acquisition to a peer-accountability framework that holds referring sellers responsible for the performance of their recruits, thereby mitigating the presence of unreliable sellers and fostering greater trust in the online shopping environment.

Page 9 of 9 | Total Record : 83

 5   6   7   8   9 

Filter by Year

2023 2026


Reset
Filter By Issues
All Issue Vol 4 No 3 (2026): JOCSTEC - September Vol 4 No 2 (2026): JOCSTEC - Mei Vol 4 No 1 (2026): JOCSTEC - Januari Vol 3 No 3 (2025): JOCSTEC - September Vol 3 No 2 (2025): JOCSTEC - Mei Vol 3 No 1 (2025): JOCSTEC - Januari Vol 2 No 3 (2024): JOCSTEC - September Vol 2 No 2 (2024): JOCSTEC - Mei Vol 2 No 1 (2024): JOCSTEC - Januari Vol 1 No 3 (2023): JOCSTEC - September Vol 1 No 2 (2023): JOCSTEC - Mei Vol 1 No 1 (2023): JOCSTEC - Januari More Issue