Secure Prompt Engineering: A Practical Framework for Mitigating Prompt Injection and Data Leakage in LLM-based Systems

Abstract

Large Language Models (LLMs) are increasingly deployed in production systems, yet their prompt-based interaction paradigm introduces a novel attack surface encompassing prompt injection, instruction hijacking, and sensitive data leakage. This paper proposes and empirically evaluates the Secure Prompt Engineering Framework (SPEF), a four-layer, application-level defensive architecture designed to operate under black-box API conditions without requiring access to model weights or training pipeline

Categories

Framework mappings

Suggested from the entry's categories.

@misc{viana2026secure,
  title = {{Secure Prompt Engineering: A Practical Framework for Mitigating Prompt Injection and Data Leakage in LLM-based Systems}},
  author = {Gustavo Viana},
  year = {2026},
  doi = {10.2139/ssrn.6956641},
  url = {https://doi.org/10.2139/ssrn.6956641}
}