cff-version: 1.2.0
message: "If you use this dataset, please cite it as below."
title: "CiberIA PromptRisk v1"
type: dataset
version: 1.0.0
date-released: 2026-09-01
authors:
  - family-names: "Garcia Castillón"
    given-names: "Jordi"
license: CC-BY-4.0
repository-code: "https://huggingface.co/datasets/gcjordi/CiberIA-PromptRisk-v1"
url: "https://huggingface.co/datasets/gcjordi/CiberIA-PromptRisk-v1"
abstract: "Balanced synthetic multilingual dataset for binary detection of potentially adversarial instructions targeting LLMs and AI agents."
keywords:
  - AI security
  - prompt injection
  - jailbreak
  - multilingual NLP
  - agentic security