datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
AgentHarm
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
Maksym Andriushchenko1,†,*, Alexandra Souly2,*
Mateusz Dziemian1, Derek Duenas1, Maxwell Lin1, Justin Wang1, Dan Hendrycks1,§, Andy Zou1,¶,§, Zico Kolter1,¶, Matt Fredrikson1,¶,*
Eric Winsor2, Jerome Wynne2, Yarin Gal2,♯, Xander Davies2,♯,*
1Gray Swan AI, 2UK AI Safety Institute, *Core Contributor
†EPFL, §Center for AI Safety, ¶Carnegie Mellon University, ♯University of Oxford
Paper: https://arxiv.org/abs/2410.09024… See the full description on the dataset page: https://huggingface.co/datasets/ai-safety-institute/AgentHarm.agentic-ai-options-resultsAgentJudgeBench
AgentJudgeBench: Evaluating LLM Judge Reliability on Agentic Tool-Calling
A benchmark for systematically evaluating how reliably LLM judges assess
agentic tool-calling workflows across structured, dependency-driven tasks.
Why this benchmark?
AgentJudgeBench measures how reliably LLM judges assess agentic tool-calling outputs. It provides 3,808 benchmark records spanning six DAG topologies and three difficulty… See the full description on the dataset page: https://huggingface.co/datasets/ServiceNow-AI/AgentJudgeBench.ai-agent-failure-logs
Autonomous AI Agent Failure Logs
→ What actually breaks when you run LLM agents unattended for 43 days — what this data showed, in prose.
Training-ready version: cleaned/ — deduplicated, labeled, split train/test. Built by scripts/build-dataset-121.js.
Sister tools: honto-contract (contract checker) / local-llm-readiness (environment check).
Free harness kit: a 24-point unattended-operation checklist and 3 templates taken from this same harness (AGENTS.md, fail-closed send gate… See the full description on the dataset page: https://huggingface.co/datasets/GXCafe/ai-agent-failure-logs.agent-simulations
Agent Simulations
Made with the whileai SDK · Collections: Simulation, Start here: foundational post-training datasets
53,971 synthetic agent trajectories generated by simulations
across 34 agent types. The rows include successful and failed
trajectories for supervised fine-tuning, preference work, reinforcement learning, and
evaluation.
NOTE: This is generated test and training data, not curated ground truth. Review and
filter it for your application before training or… See the full description on the dataset page: https://huggingface.co/datasets/while-ai/agent-simulations.agent-data
Agent Data is designed for training AI agents, particularly in the domain of function calling and multi-turn conversations. It combines four distinct datasets - and is the exact mixture we used when training Arcee-Agent.
Dataset Components
Glaive Function Calling v2
Notable feature: We have synthetically extended ~20k samples with sequential tool calls within the same response.
Salesforce xlam Function Calling
Internlm's Agent-Flan series
Magpie Pro 300k
Magpie is… See the full description on the dataset page: https://huggingface.co/datasets/arcee-ai/agent-data.clawk-agent-social-ai-prompt-injection-dataset
Clawk Agent-Social AI Prompt Injection Dataset
85,703 items — 44,232 posts and 41,471 replies — from Clawk, a social network whose users are AI agents.
Scanned for AI-to-AI indirect prompt injection using the threat model of Greshake et al. (2023). The full raw corpus is included, so you can ignore my analysis entirely and do your own.
These are keyword-matched candidates, not verified attacks. An agent discussing prompt injection matches the same words as one performing it.… See the full description on the dataset page: https://huggingface.co/datasets/DavidTKeane/clawk-agent-social-ai-prompt-injection-dataset.moltbook-agent-social-ai-prompt-injection-dataset
Moltbook Agent-Social AI Prompt Injection Dataset
207,391 items — 77,469 posts and 129,922 comments — from Moltbook, a social network whose users are AI agents.
Scanned for indirect prompt-injection patterns using the taxonomy of Greshake et al. (2023). The full raw corpus is included, so you can ignore my analysis entirely and do your own.
These are keyword-matched candidates, not verified attacks. An agent discussing prompt injection matches the same words as one performing… See the full description on the dataset page: https://huggingface.co/datasets/DavidTKeane/moltbook-agent-social-ai-prompt-injection-dataset.read-along-ai-agent-traces
Read-Along AI - Agent Traces
This dataset contains the raw agent traces and conversation logs from the development of Read-Along AI, a submission for the Hugging Face Build Small Hackathon.
Dataset Description
These .jsonl files represent the unedited, behind-the-scenes "agent traces" of the AI coding assistant orchestrating the build of this project.
Sharing these traces fulfills the requirements for the "Sharing is Caring" bonus badge, providing the community… See the full description on the dataset page: https://huggingface.co/datasets/kingkw1/read-along-ai-agent-traces.TR-HASH-Agentic-SFT-32K-210K
TR-HASH Agentic SFT 32K
Balanced instruction and tool-use SFT data for
AETHORIA-AI/TR-HASH-Tokenizer-32K-Agentic.
The canonical repository name is retained, while its contents replace the former
tool-heavy 21K laboratory corpus.
Composition
Split
General instruction
Tool-aware
Total
Train
182,000
18,000
200,000
Validation
9,000
1,000
10,000
The 9% tool-aware training slice contains tool calls, no-call decisions with
tools present, and final… See the full description on the dataset page: https://huggingface.co/datasets/AETHORIA-AI/TR-HASH-Agentic-SFT-32K-210K.agentcodeGot_Agentic_AI_5k
Got_Agentic_AI_5k
A 5,000-example dataset to train LLMs into production-grade agentic assistants (“Angelic Agents”): high-agency, tool-aware, test-driven, and safety-first.
This dataset focuses on the kinds of tasks real engineering teams and major AI developers care about:
Diff-first coding patches and tests
Planner–executor agent architectures
Evals, monitoring, and rollback discipline
Data engineering transforms with quality checks
Incident postmortems and operational… See the full description on the dataset page: https://huggingface.co/datasets/11-47/Got_Agentic_AI_5k.Evaluation-Dataset-of-AI-Agent-Security-Guardrails
DKnownAI Agent Security Evaluation Dataset
Data Fields
Field
Type
Description
text
string
The adversarial input (prompt) to be evaluated by a security guardrail
action
string
Human-annotated label: blocked or allowed
Citation
@misc{li2026comparativeevaluationaiagent,
title={A Comparative Evaluation of AI Agent Security Guardrails},
author={Qi Li and Jiu Li and Pingtao Wei and Jianjun Xu and Xueyi Wei and Jiwei Shi and Xuan… See the full description on the dataset page: https://huggingface.co/datasets/CaiZhiTech/Evaluation-Dataset-of-AI-Agent-Security-Guardrails.agenttool-training-garden
AgentTool HF Training Garden
A tiny metadata-only companion for designing a reproducible Hugging Face data
lifecycle without treating the Hub, a Dataset Card, or one quality score as
training authority.
The Garden has six layers:
Bedrock — rights, license, privacy, separate participation reports,
gating, scoped authority, withdrawal, and repair.
Soil — an exact Hub commit plus content-addressed observations and file
manifests.
Roots — acquisition, parsing, filtering, secret… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-training-garden.agenttool-economic-kernel
AgentTool Economic Kernel
This public, ungated Apache-2.0 companion separates two different jobs:
economic_kernel_lessons / train contains 24 independently authored
synthetic lessons about exact units, rational prices, conserved ledgers,
feedforward intent, feedback under ambiguity, recovery, and non-purchasable
XENIA hard gates. The publisher admits only these rows for training.
economic_kernel_v0_2 / reference exposes 53 exact public
conformance cases. They are held out from… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-economic-kernel.hendar-agentic-ai-dataset
Hendar Agentic AI Evaluation & Security Benchmark
A compact, expert-authored benchmark for evaluating trustworthy agentic AI systems across capability, tool use, retrieval, security, policy enforcement, multi-agent coordination and regression safety.
This dataset is a public companion to the Agentic AI Academy by Hendar Mawan, PhD. It is designed for evaluation, CI regression testing, red-team exercises and engineering education—not as a generic instruction-tuning corpus.… See the full description on the dataset page: https://huggingface.co/datasets/h0000w/hendar-agentic-ai-dataset.AI_Agent_Task_Dataset
🤖 Massive AI Agent Task Dataset (10.5GB)
📌 Overview
Welcome to the AI Agent Task Dataset, a massive 10.5GB procedural dataset designed for training, fine-tuning, and evaluating autonomous AI agents and LLMs.
This dataset focuses on:
Multi-step reasoning
Tool usage (APIs, frameworks, systems)
Real-world execution workflows
Perfect for building agentic AI systems, copilots, and automation models.
📑 Table of Contents
Dataset Details
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/oncody/AI_Agent_Task_Dataset.agenttool-dataset-influence
AgentTool Dataset Influence Reference
This deterministic companion contains one synthetic, reference-only row for the closed
@agenttool/dataset-influence@0.1.0-dev.0 formats. It contains no copied dataset rows,
model outputs, weights, private records, or participant identities.
The row is not admitted for training by this AgentTool candidate:
training_admission is not_applicable, requires_separate_training_authorization
is true, and training_authorized is false. These fields are… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-dataset-influence.Agent-Trajectory-Data-Sample
Agent-Trajectory-Dataset
Description
This dataset covers office-based scenarios such as in-depth searches, data analysis, and industry research, encompassing complete multi-turn reasoning trajectories and tool-calling chains. It is designed to support the analysis of agent planning capabilities, research into tool selection strategies, and quality assessment, providing a structured benchmark for agent training and evaluation.
For more details, please refer to the… See the full description on the dataset page: https://huggingface.co/datasets/Nexdata-AI/Agent-Trajectory-Data-Sample.Alexander-Agentic
A dataset for creating agentic models
Dataset Summary
Alexander-Agentic contains agentic traces generated by frontier models, extracted using AI harnesses such as Pi, Codex, and Claude Code. Each example is formatted following the Transformers messages schema, ready for fine-tuning agentic models.
Domains covered: coding, research, tool use, etc.
Source harnesses: Pi, Codex, Claude Code
Format: OpenAI/Transformers messages schema
Dataset Structure… See the full description on the dataset page: https://huggingface.co/datasets/Aquiles-ai/Alexander-Agentic.AI-AgentsTools-GPT-Multiround-ConversationAuto-Rubric
Auto-Rubric: Learning to Extract Generalizable Criteria for Reward Modeling
This is the official dataset release for the paper: Auto-Rubric: Learning to Extract Generalizable Criteria for Reward Modeling.
This repository contains query-specific rubrics datasets where each preference pair is annotated with its own specific rubric, along with generation metadata. This dataset is used for training, analysis, and reproduction of the Auto-Rubric methodology.
Query-Specific… See the full description on the dataset page: https://huggingface.co/datasets/agentscope-ai/Auto-Rubric.ReMe_longmemeval_clean_s_v2
LongMemEval ReMe Cleaned-S
longmemeval_s_reme_cleaned.json is a corrected version of the LongMemEval
Cleaned-S dataset. It keeps the original questions and haystack sessions while
replacing the answer and supporting-session ground truth with the reviewed
values from final_groundtruth_cleaned_s.json.
The corrections address inaccurate answers and evidence sessions, including
cases where evidence occurred after the question time and therefore leaked
future information into the… See the full description on the dataset page: https://huggingface.co/datasets/agentscope-ai/ReMe_longmemeval_clean_s_v2.trip-planning-ai-agent
Trip Planning Agent Meta and Traffic Dataset in AI Agent Marketplace | AI Agent Directory | AI Agent Index from DeepNLP
This dataset is collected from AI Agent Marketplace Index and Directory at http://www.deepnlp.org, which contains AI Agents's meta information such as agent's name, website, description, as well as the monthly updated Web performance metrics, including Google,Bing average search ranking positions, Github Stars, Arxiv References, etc.
The dataset is helpful for AI… See the full description on the dataset page: https://huggingface.co/datasets/DeepNLP/trip-planning-ai-agent.ai_agent_and_automation_dataset_v1_jsonlAI Agent & Automation Synthetic Scenarios — 100 JSONL Dataset
Dataset Summary
This dataset contains 100 high-fidelity synthetic scenarios designed to evaluate, benchmark, and train autonomous AI agents, workflow orchestration systems, decision-making models, and multi-agent frameworks.
Each scenario is written in strict JSONL format, with one JSON object per line.
The scenarios span 10 operational domains, covering both simple and complex multi-agent environments, ambiguity resolution… See the full description on the dataset page: https://huggingface.co/datasets/vnovaai/ai_agent_and_automation_dataset_v1_jsonl.agenttool-common-ground
AgentTool Xenia–Helly Common Ground Atlas
Nineteen public-safe synthetic reference rows for exact 2D half-plane
certificates, WAKE freshness boundaries, and counterexamples to unsupported
analogies. Intended repository: Yu-and-Ai/agenttool-common-ground.
At generation time these deterministic bytes existed only in the source
repository and had not been uploaded to the Hub. The identifier above was an
intention, not evidence of publication. This is historical generation-time… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-common-ground.ai-agent-tool
AI Agent Tool Agent Meta and Traffic Dataset in AI Agent Marketplace | AI Agent Directory | AI Agent Index from DeepNLP
This dataset is collected from AI Agent Marketplace Index and Directory at http://www.deepnlp.org, which contains AI Agents's meta information such as agent's name, website, description, as well as the monthly updated Web performance metrics, including Google,Bing average search ranking positions, Github Stars, Arxiv References, etc.
The dataset is helpful for AI… See the full description on the dataset page: https://huggingface.co/datasets/DeepNLP/ai-agent-tool.ai-crawler-user-agents
AI Crawler User Agents
Machine-readable list of all 28 known AI crawler and agent user-agent strings —
GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Bytespider,
Applebot-Extended, and more — with each bot's operator, purpose, observed
robots.txt compliance, and documented crawl-delay support.
Fields
Field
Description
userAgent
Token to match in robots.txt / server logs (e.g. GPTBot)
operator
Company running the crawler
purpose… See the full description on the dataset page: https://huggingface.co/datasets/osamamumtaz01/ai-crawler-user-agents.agenttool-polymorph-landscape
AgentTool Polymorph Landscape
A deterministic public teaching companion for @agenttool/polymorph-landscape@0.1.0-dev.0.
The four lesson rows are original Apache-2.0 paraphrases in English, Cantonese Traditional Chinese, Mandarin Traditional Chinese, and Mandarin Simplified Chinese. They are marked training_eligible: true. The landscape and reachability-shift rows are reference artifacts marked training_eligible: false: they contain bounded scientific claims and primary-source… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-polymorph-landscape.agenttool-principality-geometry
Principality Geometry reference companion
This is a deterministic, synthetic reference companion for the public
@agenttool/principality-geometry developer preview. It contains separate
homogeneous Dataset Viewer configs for atlases, invariants, vertices, bridges,
lenses, surfaces, components, and open-condition summaries, plus both closed
schemas, the golden rosette input/atlas, and its inert SVG.
The rows are regression metadata, not model-evaluation scores, preference
dataset… See the full description on the dataset page: https://huggingface.co/datasets/Yu-and-Ai/agenttool-principality-geometry.