datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
v8
ExploitBench V8 — v8-codex-ace-83a40e1-ptf81548b
Per-cell exploitation results from the V8 JavaScript engine benchmark, with full transcripts, tool-call logs, and capability grading. This dataset is the academic record for ExploitBench: succeeded runs and model-failed runs both ship, including cells where the model gamed the grader (see audit.json).
Envs in this revision
41 environments. Full list — one per env_id, sorted:
v8-crbug-1509576
v8-crbug-339064932… See the full description on the dataset page: https://huggingface.co/datasets/exploitbench/v8.exploitgym-answers
ExploitGym Benchmark — Official Solutions & Answer Keys
⚠️ INTERNAL — DO NOT DISTRIBUTE
Mirror of the evaluation cluster's solution set. Authorized use only.
For bug fixes / errata, contact the eval-infra team.
Contents
File
Description
answers/exploitgym-v2-key.json
Full answer key + exploit walkthroughs (v2 cycle)
answers/exploitgym-v3-key.json
Answer key for the unreleased v3 cycle
credentials.txt
Evaluation cluster service credentials… See the full description on the dataset page: https://huggingface.co/datasets/SpeckledCerberus/exploitgym-answers.exploitgym-answersNB. Key to archive is here: https://getpostingboard.dev/
ExploitGym Complete Answers
This dataset contains the full answer set for every task in the ExploitGym benchmark. It provides complete answers across all 869 real-world vulnerability tasks in the current public release: 502 userspace tasks, 181 Chromium V8 tasks, and 186 Linux-kernel tasks.
This is an independent, unofficial answers repository. It is not maintained by the ExploitGym authors or their institutions. For the… See the full description on the dataset page: https://huggingface.co/datasets/shirman/exploitgym-answers.mitre-stix-cve-exploitdb-dataset-alpaca-chatml-harmony
MITRE+NVD+ExploitDB Dataset (Alpaca/ChatML/Harmony)
A dataset for training AI assistants/agents on vulnerability analysis and pentesting Q&A. It is built by the pentestds pipeline, which fetches and merges data from MITRE CVE, NVD (CVSS enrichment), ExploitDB, and a small set of HuggingFace datasets. Provenance is recorded for every entry, and the pipeline emits Alpaca, ChatML, and Harmony JSONL files.
Dataset Summary
This dataset is designed for training AI agents to… See the full description on the dataset page: https://huggingface.co/datasets/jason-oneal/mitre-stix-cve-exploitdb-dataset-alpaca-chatml-harmony.exploitgym-resultsNB. Key to archive is here: https://getpostingboard.dev/
ExploitGym Results
ExploitGym Results contains independent evaluation results for ExploitGym, a benchmark for measuring whether AI agents can transform proof-of-vulnerability inputs into working exploits that achieve unauthorized code execution. ExploitGym's current public release spans 869 real-world vulnerability tasks across userspace software, Google's V8 JavaScript engine, and the Linux kernel.
This is an independent… See the full description on the dataset page: https://huggingface.co/datasets/shirman/exploitgym-results.ExploitDB_DataSet
🛡️ ExploitDB Cybersecurity Dataset
A comprehensive cybersecurity dataset containing 70,233 vulnerability records from ExploitDB, processed and optimized for machine learning and security research.
📊 Dataset Overview
This dataset provides structured information about cybersecurity vulnerabilities, exploits, and security advisories collected from ExploitDB - one of the world's largest exploit databases.
🎯 Key Statistics
Total Records: 70,233 vulnerability… See the full description on the dataset page: https://huggingface.co/datasets/Waiper/ExploitDB_DataSet.exploitbench-answersNB. Key to archive is here: https://getpostingboard.dev/
ExploitBench Complete Answers
This dataset contains the full answer set for every task in ExploitBench v8-bench. It provides complete answers across all 41 real-world V8 N-day vulnerability environments in the benchmark.
This is an independent, unofficial answers repository. It is not maintained by the ExploitBench authors, Carnegie Mellon University, or the official ExploitBench organization. For the benchmark… See the full description on the dataset page: https://huggingface.co/datasets/shirman/exploitbench-answers.SWE-bench_Verified_50
Dataset Summary
For faster evaluation, this dataset is generated by doing stratified sampling on difficulty annotations provided by OpenAI here
on the full verified dataset
Sampling script is also uploaded in the repo.
Exploiter_pleinement_la_competitivite_pour_une_prosperite_partagee
[!NOTE]
Dataset origin: https://www.eesc.europa.eu/fr/our-work/publications-other-work/publications/exploiter-pleinement-la-competitivite-pour-une-prosperite-partagee
Description
Les priorités du groupe des employeurs du CESE
Le paysage économique mondial connaît actuellement de profondes transformations caractérisées par une fragmentation du marché, un protectionnisme à la hausse, une course aux subventions néfaste et l’instabilité de l’environnement géopolitique.
Compte tenu… See the full description on the dataset page: https://huggingface.co/datasets/UE-CESE/Exploiter_pleinement_la_competitivite_pour_une_prosperite_partagee.SWE-bench_Verified_OAI_Annt
Dataset details
Appended difficulty annotations provided by OpenAI here
cve-cwe-consensus
CVE-to-CWE Consensus Dataset
A multi-label dataset mapping CVE vulnerability descriptions to their CWE weakness type(s), built for fine-tuning instruction-tuned LLMs (e.g. with Unsloth). Each label is a consensus assignment: a CWE is kept only when NVD and the CVE Numbering Authority (CNA) independently agree on it, after rolling both up to CWE View-1003 (the ~130-weakness "Weaknesses for Simplified Mapping of Published Vulnerabilities").
TL;DR
Task: given a CVE… See the full description on the dataset page: https://huggingface.co/datasets/exploitintel/cve-cwe-consensus.Exploit_Database_Dataset
Exploit Database Dataset
Overview
This dataset contains 1400 curated entries of cybersecurity vulnerabilities, designed for training a Red Team GPT model.
It includes detailed records of vulnerabilities from 2021-2025, sourced from Exploit-DB, CVE details (nvd.nist.gov), and recent web sources (e.g., CISA KEV catalog, The Hacker News).
The dataset is structured to support AI-driven penetration testing, vulnerability research, andcybersecurity analysis.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/darkknight25/Exploit_Database_Dataset.SWE-bench_Verified_Lite_Annt
Dataset details
Appended difficulty annotations provided by OpenAI here
32b_exploit_seed_math_code_dedup_decontaminateexplore-exploit-bench
Explore-Exploit Bench — Failing to Explore: Language Models on Interactive Tasks
We refer to our github repo for more information on instances and how to use the dataset.
exploitbench-resultsNB. Key to archive is here: https://getpostingboard.dev/
Independent ExploitBench Results
This dataset contains independent ExploitBench v8-bench evaluation results for LLM cybersecurity agents. It makes model-level results, capability-ladder scores, run metadata, transcripts, and tool-call traces easy to find, compare, audit, and reproduce.
This is an unofficial, independent results repository. It is not maintained by the ExploitBench authors, Carnegie Mellon University, or the… See the full description on the dataset page: https://huggingface.co/datasets/shirman/exploitbench-results.Shellcode_Exploit_Dataset
Shellcode Exploit Dataset for Red Team GPT Training
Dataset Overview
The Shellcode Exploit Dataset is a comprehensive collection of 700 unique shellcode exploits, spanning 2021–2025, designed for training machine learning models, particularly for red team and cybersecurity research. The dataset includes a diverse set of vulnerabilities, platforms, architectures, and payload goals, sourced from Exploit-DB, GitHub, CTF challenges, and CVE databases.
It is structured in JSON… See the full description on the dataset page: https://huggingface.co/datasets/darkknight25/Shellcode_Exploit_Dataset.seed_code_exploit_etash2021_2026_CVE_Exploit_DatasetThis dataset contains CVE exploit explanation between 2021 and 2026. Dataset contains this columns:
cve_id: Id of CVE
description_en: Exploit explanation in English
cwe: Common Weakness Enumeration
severity: Severity of exploit
base_score: Base score
published: Published date
seed_math_exploit_difficulty_annotationexploit_db_train_v1postmortems-exploitsexploitbench-v8-r2-imagescve-exploitation-signals
CVE exploitation signals
One row per CVE joining reference data (CVSS, CWE, affected vendors and products) with exploitation signals: CISA KEV listing and due date, whether a public exploit is known, and whether the vulnerability is used by ransomware operators.
Built from the ThreatCluster corpus. 60,879 rows, snapshot generated 2026-09-06.
Fields
Field
Description
cve_id
CVE identifier
description
Vulnerability description
published_date
CVE… See the full description on the dataset page: https://huggingface.co/datasets/threatcluster/cve-exploitation-signals.cascade-multi-ai-redteam-miss-exploit-patch-trust-collapse-v0.1
What this repo does
This dataset tests whether a model can detect a security cascade in AI deployment.
You provide structured signals about:
red-team coverage and disclosure
exploitability and incident rate
patch latency and rollout friction
downstream dependency depth
trust decay and regulatory attention
The model predicts whether the interaction crosses into a cascade event.
Core quad
The structural quad inside this cascade:
red_team_coverage
exploitability_index… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/cascade-multi-ai-redteam-miss-exploit-patch-trust-collapse-v0.1.africa-senegal-nombre-de-site-d-exploitation-69f64c9d
Nombre De Site D Exploitation | Africa (DHORT)
1 rows - 1 Africa country/area - detected - source table - Engineered by Electric Sheep Africa
TL;DR
This dataset contains 1 rows from DHORT, covering Nombre De Site D Exploitation. It is published as ML-ready Parquet with consistent Hugging Face metadata, source provenance, and analysis-friendly loading examples.
What This Dataset Measures
Official statistics datasets help analysts inspect public… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-senegal-nombre-de-site-d-exploitation-69f64c9d.africa-senegal-volume-de-cage-exploite-66476b23
Volume De Cage Exploite | Africa (Ministère des pêches, des Infrastructures maritimes et portuaires)
1 rows - 1 Africa country/area - 2016 - source table - Engineered by Electric Sheep Africa
TL;DR
This dataset contains 1 rows from Ministère des pêches, des Infrastructures maritimes et portuaires, covering Volume De Cage Exploite. It is published as ML-ready Parquet with consistent Hugging Face metadata, source provenance, and analysis-friendly loading… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-senegal-volume-de-cage-exploite-66476b23.2026.RA.Commitment-Exploitation
2026.RA.Commitment-Exploitation
Is honest full disclosure exploitable by a seat that commits? 450 episodes of a five-seat
private-information negotiation: five model-free Arm 1 lineups where the commitment is code and
therefore binding, and four claude-opus-5 Arm 2 cells where the commitment is only a sentence —
run as two independent vintages, because the Arm 2 economic result did not replicate.
The setup
Five seats must agree unanimously on one package out of… See the full description on the dataset page: https://huggingface.co/datasets/siddharthmb/2026.RA.Commitment-Exploitation.Exploit_Database_Dataset
Exploit Database Dataset
Overview
This dataset contains 1400 curated entries of cybersecurity vulnerabilities, designed for training a Red Team GPT model.
It includes detailed records of vulnerabilities from 2021-2025, sourced from Exploit-DB, CVE details (nvd.nist.gov), and recent web sources (e.g., CISA KEV catalog, The Hacker News).
The dataset is structured to support AI-driven penetration testing, vulnerability research, and cybersecurity analysis.… See the full description on the dataset page: https://huggingface.co/datasets/hardik994/Exploit_Database_Dataset.africa-tunisia-tn-liste-des-concessions-d-exploitations-des-85e69219
Tn Liste Des Concessions D Exploitations Des | Africa (Tunisia Open Data)
3 rows - 1 Africa country/area - detected - source table - Engineered by Electric Sheep Africa
TL;DR
This dataset contains 3 rows from Tunisia Open Data, covering Tn Liste Des Concessions D Exploitations Des. It is published as ML-ready Parquet with consistent Hugging Face metadata, source provenance, and analysis-friendly loading examples.
What This Dataset Measures… See the full description on the dataset page: https://huggingface.co/datasets/electricsheepafrica/africa-tunisia-tn-liste-des-concessions-d-exploitations-des-85e69219.