GenAI & Agentic AI Security Incidents — 7,725 real-world & research incidents

Research
1

:bar_chart: Just released: GenAI & Agentic AI Security Incidents

A dataset of 7,725 real-world and research incidents covering:

Prompt injection, jailbreaks, data exfiltration, deepfakes, agent hijacking, AI-enabled harms, and more.

Each incident is mapped to:

OWASP LLM Top 10 2025
OWASP Agentic AI Security Initiative Top 10
NIST AI RMF
MITRE ATLAS techniques and tactics

from datasets import load_dataset

ds = load_dataset("emmanuelgjr/genai-incidents", split="train")

# Example: filter for prompt injection incidents
prompt_injection = ds.filter(
    lambda r: "LLM01" in (r["owasp_llm"] or [])
)

Each entry includes a quality_tier field — curated, reviewed, or auto — so researchers and practitioners can filter by vetting level.

Licensed under CC-BY-4.0 and citable with a DOI.

Feedback, issues, and PRs are very welcome. :folded_hands:

:backhand_index_pointing_right: https://huggingface.co/datasets/emmanuelgjr/genai-incidents

1 Like