Agentic AI Research Radar

License: MIT PRs Welcome Maintenance GitHub stars

300+ curated agentic AI papers, benchmarks, frameworks, and real-world systems across 24 domains.

Start with Frameworks | Explore Security | Browse Research | Contribute a Resource

Why This Repo

Agentic AI moves fast, and useful work is scattered across papers, code repositories, benchmark releases, technical posts, and domain-specific case studies. This repo tracks high-signal work in one place so builders and researchers can quickly find:

  • frameworks and orchestration patterns for building agents
  • evaluation, testing, and benchmark resources
  • security, safety, and governance research
  • domain applications in healthcare, Physical AI/robotics, software engineering, IoT, education, and more

If this saves you research time, starring the repo helps more builders discover it.

Suggested Paths

Path Follow these domains
Core AI & Cognition Foundation Models, Reasoning, Planning, Vision, Recommendation
Engineering, DevTools & Frameworks Software Engineering, Testing, Framework, Automation
Applied Industry & Civic Domains Healthcare, Security, OT/IoT, Smart Cities, Transportation, Supply Chain
Hardware, Physical Systems & Networks Physical AI/Robotics, Hardware Design, Design, Networking, Blockchain
Specialized Ecosystems, Learning & Ethics Research, Education, Gaming, Ethics and Safety

Domain Index

Domain Researches Focus
🔬 Research 64 AI scientists, deep research, autonomous discovery
⚙️ Framework 53 Agent runtimes, memory, tools, orchestration
🛡️ Security 50 Prompt injection, red teaming, cyber agents, defenses
🧠 Reasoning 25 Inference, deliberation, cognitive architectures
🧪 Testing 19 Benchmarks, harnesses, evaluation methodology
💊 Healthcare 15 Medical agents, diagnosis, pharma, biomedical research
💻 Software Engineering 15 Coding agents, debugging, refactoring, code review
⚡ Automation 13 Process automation and workflow optimization
📡 Networking 9 Self-healing networks and routing
👁️ Vision 7 Visual agents and multimodal grounding
🤖 Physical AI/Robotics 7 Physical AI, embodied agents, and robot systems
🤖 Foundation Models 6 Agentic capabilities in foundation models
🔧 Hardware Design 6 EDA, chip design, verification
⚖️ Ethics and Safety 5 Alignment, fairness, trust, governance
🏭 OT/IoT 5 Industrial systems and device control
🎯 Recommendation 5 Personalization and proactive discovery
🌆 Smart Cities 5 Urban operations and civic systems
⛓️ Blockchain 4 Decentralized agents and secure transactions
🎮 Gaming 4 NPCs, simulation, procedural content
📐 Design 3 Generative and collaborative design
🚦 Transportation 2 Autonomous vehicles and fleet management
📋 Planning 2 Multi-step decision making and goal assignment
🎓 Education 1 Tutoring and personalized learning
📦 Supply Chain 1 Routing, inventory, logistics optimization

Curation Standard

The repo is moving toward GitHub-native, canonical resource links:

  • Primary links should point to the paper, code repository, benchmark, docs, or project page.
  • LinkedIn/X posts are useful as commentary, but should be secondary links.
  • Titles should be plain readable text.
  • Resources should be placed in the most relevant domain and avoid duplicates.

See CONTRIBUTING.md for the full contribution checklist. See CURATION.md for the detailed resource quality bar and link-priority rules.

Author

Mahmoud Rabie