Trustworthy Retrieval-Grounded AI for Technical Learning

Linux manual pages (man pages) provide comprehensive documentation for Unix-based systems, yet their usage remains limited by poor discoverability and steep learning curves, particularly for novice users. This paper proposes a Retrieval-Augmented Generation (RAG) based assistant that enables natural language querying over the full corpus of Linux man pages. The system combines structured data extraction, semantic indexing, and an open-source large language model to provide accurate, context-aware responses without retraining the base model. Linux is essential for software developers, cyber security specialists, students. The proposed RAG-enhanced LLM system for Linux learners can be used by academics to support student learning. It has been tested by learners and academics and is a relevant case study for the integration of AI tools in higher education. Beyond documentation access, the system functions as an AI-supported learning scaffold that promotes independent problem-solving, reduces cognitive overload, and supports the development of professional digital fluency skills central to future-facing computing education.