🚀 Getting Started | 🎥 Meeting Recordings | 📧 Google Group
- Every Friday
- 🕕 18:00–19:00 CET | 🕔 17:00–18:00 London | 🕘 09:00–10:00 PST
- 🔗 Google Meet | 📅 Google Calendar | 📥 .ics Download
Not in CNCF Slack? Join here first
Join the #agentic-community channel
A collaborative open initiative by member organizations to bridge AI research and enterprise reality through practical agent applications
Agentic AI has surpassed our understanding of designing scalable, trustworthy, and maintainable agent systems. The Agentic Community initiative brings together practitioners to share, debate, and refine design patterns, evaluation strategies, and capability-first approaches for architecting and developing agentic systems. Our goal is to help organizations transition from agent experimentations to successful agent deployments.
We're a collaborative initiative focused on practical agent applications using existing, open source tools, cloud services, frameworks and standards. As this field is still developing, we may find an area that is under served for us to contribute through code contributions or thought leadership. We aim to address the following:
🤖 Agent Autonomy
- Decision framework for rules based vs LLM augmented vs Agent implementation
- Autonomy levels and risk assessment
- Real-world case studies and patterns
- Implementation guidelines by autonomy tier
🔗 Multi-Agent Architecture
- Single vs multi-agent decision criteria
- Model selection strategies (fixed vs dynamic)
- Self-organizing agent patterns
- Orchestration and coordination considerations and patterns
Future-Ready Design
- Protocol-agnostic architecture patterns
- Component isolation strategies
- Adaptation paths between protocols (MCP/A2A/ACP)
- Migration strategies for emerging technologies
🎛️ Enterprise Controls
- Transparency requirements by use case
- Audit mechanisms and logging standards
- Control frameworks by trust level
- Compliance considerations
🔐 Security & Trust
- Security-first design principles
- Human oversight integration
- Value alignment verification
- Risk mitigation strategies
Standards & Interoperability
- Shape existing open source projects and standards critical for security, transparency, and interoperability
📦 Deliverables:
- An in-depth understanding of the agentic ecosystem leading to informed approaches in the form of white papers, blogs, patterns, and best-practices regarding agent development, interoperability, operations, and security.
- An architecture for building, testing, and deploying agents
- Collaborative and open problem-solving approach through shared experiences
🤝 Join the Community
- Contribute Your Experience: Share successful (and unsuccessful) agent implementations to help others learn
- Collaborate on Patterns: Join working groups focused on specific industries or use cases
- Shape the Ecosystem: Guide the evolution of tools and standards that enable secure, transparent, and interoperable agent systems
🏁 Getting Started
- Read our charter.md to understand our principles
- Read our onboarding guide, attend a meeting, and introduce yourself
- Join discussions to connect with practitioners and propose use cases or patterns
🧠 Your experience and perspective can help shape the future of agentic AI—collaboratively, transparently, and openly.
This project is licensed under the Apache-2.0 License.