Publications
9 papers on agent security, prompt-injection defence, edge verification and satellite networking. Each page carries its full abstract and a link to the PDF, is marked up so that academic indexers can read the citation, and states whether the work is independently checkable.
- Edge-Native Semantic Firewall for Autonomous LLM Agents: A Structured Chain-of-Thought Verification FrameworkacceptedAccepted; camera-ready in progressartifact: public
- PREBAS: Preemptive Bandwidth Scaling for WebRTC in LEO Satellite Networks Using Lightweight Neural Predictionaccepted2026 IEEE RTC — Chicago, USAartifact: none released
- A Hybrid Causal-Generative Multi-Agent Framework for Secure Autonomous Mechatronics: Mitigating Real-Time Prompt Injections in Edge-Deployed Disaster Recovery NetworksacceptedAIMS 2026 — 23–24 July 2026; IEEE Xploreartifact: none released
- Evaluating Google's Open Knowledge Format (OKF v0.1) for Multi-Agent Threat Intelligence: A Proof-of-Concept in Digital ForensicsacceptedIC-AICSE 2026 — Everest Engineering College, Nepal; 7 October 2026artifact: none released
- Evaluating Wi-Fi Sensing Attack Vectors: AI-Driven Covert Threat Detection and Deterministic Mitigation in Mobile and Edge EcosystemsmanuscriptManuscriptartifact: none released
- Direct-to-Device LEO Satellite Communication: A System Architecture for Universal 5G NTN Connectivity with Focus on NepalpresentedSpace Con 2026no experiment asserted
- Project Swayam: The Unbreakable Municipality — Building Nepal's First Fully Offline, Sovereign AI Governance SystempresentedMunicipal AI Governance Conference 2026, Budhanilkantha Municipality, Nepalartifact: none released
- EmbodiedOS: Embodied Intelligence — A Practical Software Architecture for Real-Time Robotic Decision MakingmanuscriptTechnical report, v0.1.0artifact: none released
- Feasibility Study for Establishing Nepal as a Data Center Hub of South AsiamanuscriptManuscriptartifact: none released
Authors are listed in the order printed on each paper. Where a paper is co-authored, the byline says so. PDFs are the authors’ own copies hosted here so that each paper has a stable URL that can be cited. “Artifact” says what you can check for yourself: one paper ships its code, corpus and raw model outputs and re-derives its own numbers in CI. Of the others, several report measurements with no released artifact and several assert no experiment at all — architecture and feasibility work — and each page says which it is rather than presenting either as reproducible.
Acceptance status is the author’s own claim and is not independently checkable from a public source. Where it rests on a notification rather than a public record, the paper’s page says so.