mlsanityData to 7 Oct 2026

The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI Systems

arXiv 2602.17753 ↗First posted 19 Feb 2026cs.CY cs.AIcited by 27

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

Agentic AI systems are increasingly capable of performing professional and personal tasks with limited human involvement. However, tracking these developments is difficult because the AI agent ecosystem is complex, rapidly evolving, and inconsistently documented, posing obstacles to both researchers and policymakers. To address these challenges, this paper presents the 2025 AI Agent Index. The Index documents information regarding the origins, design, capabilities, ecosystem, and safety features of 30 state-of-the-art AI agents based on publicly available information and email correspondence with developers. In addition to documenting information about individual agents, the Index illuminates broader trends in the development of agents, their capabilities, and the level of transparency of developers. Notably, we find different transparency levels among agent developers and observe that most developers share little information about safety, evaluations, and societal impacts. The 2025 AI Agent Index is available online at https://aiagentindex.mit.edu

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 54 of its 140 references that are in the map. First posted on arXiv 19 Feb 2026.

Builds on

54 of 140 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents26 Oct
  2. 2025Security Challenges in AI Agent Deployment: Insights from a Large Scale Public Competition28 Jul
  3. 2025OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety8 Jul
  4. 2025How to evaluate control measures for LLM agents? A trajectory from today to superintelligence7 Apr
  5. 2025The AI Agent Index3 Feb
  6. 2024Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents11 Oct
  7. 2024LLM Agents can Autonomously Hack Websites6 Feb
  8. 2023Identifying the Risks of LM Agents with an LM-Emulated Sandbox25 Sep

Other references · 46

  1. 2018Datasheets for Datasetscited by 1,371
  2. 2018Model Cards for Model Reportingcited by 1,068
  3. 2025The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Searchcited by 375
  4. 2023An Overview of Catastrophic AI Riskscited by 214
  5. 2025GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Taskscited by 142
  6. 2023Harms from Increasingly Agentic Algorithmic Systemscited by 134
  7. 2024Evaluating Frontier Models for Dangerous Capabilitiescited by 121
  8. 2024The Ethics of Advanced AI Assistantscited by 103
  9. 2020Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Databasecited by 95
  10. 2025A Survey of WebAgents: Towards Next-Generation AI Agents for Web Automation with Large Foundation Modelscited by 87
  11. 2025Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluationcited by 83
  12. 2024Visibility into AI Agentscited by 77
34 more
  1. 2025A Survey of AI Agent Protocolscited by 77
  2. 2025Agentic AI for Scientific Discovery: A Survey of Progress, Challenges, and Future Directionscited by 74
  3. 2025From AI for Science to Agentic Science: A Survey on Autonomous Scientific Discoverycited by 74
  4. 2025International AI Safety Reportcited by 64
  5. 2024Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?cited by 56
  6. 2025Measurement to Meaning: A Validity-Centered Framework for AI Evaluationcited by 54
  7. 2024Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating Worldcited by 49
  8. 2024Safety Cases: How to Justify the Safety of Advanced AI Systemscited by 48
  9. 2018Interpretable to Whom? A Role-based Model for Analyzing Interpretable Machine Learning Systemscited by 47
  10. 2022Accountability in an Algorithmic Society: Relationality, Responsibility, and Robustness in Machine Learningcited by 47
  11. 2025Levels of Autonomy for AI Agentscited by 45
  12. 2025Infrastructure for AI Agentscited by 40
  13. 2024AI and the Problem of Knowledge Collapsecited by 34
  14. 2023Ecosystem Graphs: The Social Footprint of Foundation Modelscited by 31
  15. 2025Characterizing AI Agents for Alignment and Governancecited by 30
  16. 2022Discovering Agentscited by 28
  17. 2025BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agentscited by 23
  18. 2026OpenClaw Agents on Moltbook: Risky Instruction Sharing and Norm Enforcement in an Agent-Only Social Networkcited by 19
  19. 2022Reward Reports for Reinforcement Learningcited by 18
  20. 2024From "AI" to Probabilistic Automation: How Does Anthropomorphization of Technical Systems Descriptions Influence Trust?cited by 18
  21. 2024The 2024 Foundation Model Transparency Indexcited by 18
  22. 2025Remote Labor Index: Measuring AI Automation of Remote Workcited by 18
  23. 2025An Economy of AI Agentscited by 17
  24. 2025OCCULT: Evaluating Large Language Models for Offensive Cyber Operation Capabilitiescited by 12
  25. 2022System Cards for AI-Based Decision-Making for Public Policycited by 11
  26. 2025AI Agents and the Lawcited by 9
  27. 2025The 2025 Foundation Model Transparency Indexcited by 9
  28. 2026Legal Alignment for Safe and Ethical AIcited by 9
  29. 2024Improving governance outcomes through AI documentation: Bridging theory and practicecited by 8
  30. 2025AI Behind Closed Doors: a Primer on The Governance of Internal Deploymentcited by 8
  31. 2025Build Agent Advocates, Not Platform Agentscited by 8
  32. 2025EvalCards: A Framework for Standardized Evaluation Reportingcited by 4
  33. 2025Cybersecurity AI: The Dangerous Gap Between Automation and Autonomycited by 3
  34. 2025A New Digital Divide? Coder Worldviews, the Slop Economy, and Democracy in the Age of AIcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 27.

27 papers.

Sort:

By area:

  1. 2026Dive into Claude Code: The Design Space of Today's and Future AI Agent Systemscited by 33
  2. 2026A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Frameworkcited by 9
  3. 2026AutoResearch AI: Towards AI-Powered Research Automation for Scientific Discoverycited by 9
  4. 2026Agentic AI in the Software Development Lifecycle: Architecture, Empirical Evidence, and the Reshaping of Software Engineeringcited by 6
  5. 2026Safety in Self-Evolving LLM Agent Systems: Threats, Amplification, and Case Studiescited by 6
  6. 2026HDP: A Lightweight Cryptographic Protocol for Human Delegation Provenance in Agentic AI Systemscited by 5
  7. 2026Sovereign Agentic Loops: Decoupling AI Reasoning from Execution in Real-World Systemscited by 5
  8. 2026The Moltbook Observatory Archive: an incremental dataset of agent-only social network activitycited by 3
  9. 2026Responsible Agentic AI Requires Explicit Provenancecited by 3
  10. 2026Dissociative Identity: Language Model Agents Lack Grounding for Reputation Mechanismscited by 3
  11. 2026Synergy: A Next-Generation General-Purpose Agent for Open Agentic Webcited by 2
  12. 2026Will the Agent Recuse, and Will It Stop? Measuring LLM-Agent Compliance with In-Band Governance Signals at the Access Door and Mid-Flightcited by 2
15 more
  1. 2026Misinformation Propagation in Benign Multi-Agent Systemscited by 2
  2. 2026OmnilingualGAIA2: Evaluating the Multilingual Gap in Frontier AI Agentscited by 2
  3. 2026A Methodology for Auditable Trustworthiness Levels in AI Lifecycle Governancecited by 1
  4. 2026Graphectory Viewer: A Tool for Process-Centric Analysis of Agentic Software Trajectoriescited by 1
  5. 2026Questionnaire Responses Do not Capture the Safety of AI Agentscited by 0
  6. 2026Ambient Persuasion in a Deployed AI Agent: Unauthorized Escalation Following Routine Non-Adversarial Content Exposurecited by 0
  7. 2026RiskNet: A large-scale dataset of AI risk incidents from news with alignment and multi-dimensional annotationscited by 0
  8. 2026The Agentic Web Requires New Normative Infrastructurecited by 0
  9. 2026Why Trust Your Agent? Empirical Security Gains from TRiSM-Guided Agentic Workflows in Healthcarecited by 0
  10. 2026Specifying the Delegated-Autonomy Boundary: Requirements Engineering for Agentic AIcited by 0
  11. 2026Operational Identity: A Finite Audit of Declared and Implemented Rules of Samenesscited by 0
  12. 2026Distributing Security Controls Through Harness Engineeringcited by 0
  13. 2026Regret Dominates Surprise: Design-Time Requirements Engineering for Agentic-AI Safetycited by 0
  14. 2026The Agentic Company OS: Substrate Inversion for Sustained Enterprise Agent Deploymentcited by 0
  15. 2026Auditing Agent Actions through Query-Conditioned Attributioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025AI Agents: Evolution, Architecture, and Real-World Applications
  2. 2026Towards Agentic AI Governance: A Preliminary Assessment
  3. 2026Taxonomy and Consistency Analysis of Safety Benchmarks for AI Agents
  4. 2025Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges
  5. 2026Beyond single-channel agentic benchmarking
  6. 2024The Landscape of Emerging AI Agent Architectures for Reasoning, Planning, and Tool Calling: A Survey
  7. 2025AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges
  8. 2026The Agentic Automation Canvas: a structured framework for agentic AI project design