Several open-source standards have emerged for AI agents and agentic frameworks, aiming to improve interoperability and standardization in the rapidly evolving field of AI. Some of the key standards include:
Some of the key standards include:
- OASF (Open Agentic Schema Framework): Launched in early 2025, OASF provides standardized schemas for defining AI agent capabilities, interactions, and metadata. It offers:
• Structured ways to describe agent attributes and relationships
• Development tools and schema validation
• A foundation for interoperable AI agent systems - ACP (Agent Connect Protocol): Part of the AGNTCY initiative, ACP facilitates connections and interactions between AI agents, enabling more effective communication and collaboration.
- AGNTCY: An open-source collective launched in early 2025 to create standards for agent communication and interoperability. It includes:
• Open agent schema framework
• Agent directory for discovering compatible agents
• Agent connect protocol for cross-framework communication - LangChain’s Agent Protocol: Introduced in late 2024, this protocol allows LangChain agents to communicate with those created using other frameworks like AutoGen and CrewAI.
- MCP (Model Context Protocol): Developed by Anthropic in late 2024, MCP aims to standardize connections between AI models, tools, and data sources.
Model Context Protocol and its impact on Telecom
The telecommunications landscape is undergoing a fundamental transformation. While Telcos are currently investing significantly in Network APIs to expose 5G capabilities, edge computing, and network services to developers, this approach may soon become insufficient. As developers increasingly work with AI agents rather than traditional applications, Telcos must evolve beyond conventional APIs toward MCP-based connectivity layers that are AI-native, context-aware, and designed for autonomous AI applications.
This represents not merely an API evolution but a complete paradigm shift in integration philosophy. AI-focused companies are rapidly implementing MCP as the foundation for their agent ecosystems.
Telcos that fail to adapt, risk becoming relegated to passive infrastructure providers instead of active enablers of AI-driven connectivity.
Success will not be determined by who offers superior network APIs, but by who recognizes that we’re moving from an API-first to an AI-agent-first future.
MCP is an emerging framework that creates an abstraction layer for API-agent interactions, establishing a persistent environment for rules, permissions, and context across various integrations. Unlike traditional development requiring hardcoded API calls, AI agents can independently navigate MCP servers to understand and utilize capabilities without rigid predefined calls. This distinction is crucial: while APIs expose functions, MCP provides capabilities, prompts, and instructions that agents can interpret with flexibility. Traditional APIs demand manual integration, whereas MCP enables AI agents to transition seamlessly between different systems without constant reconfiguration.
How does the Open Agentic Schema Framework (OASF) enhance interoperatbility among AI agents
The Open Agentic Schema Framework (OASF) significantly enhances interoperability among AI agents by addressing key technical and structural challenges in multi-agent ecosystems. Here’s how it achieves this:
Standardized Data Representation – OASF provides uniform schemas for defining:
- Agent capabilities and metadata
- Data exchange formats (inputs/outputs)
- Validation rules for structured data
This eliminates inconsistencies caused by proprietary data models, enabling agents from different vendors to interpret shared data accurately.
Unified Discovery Mechanism – The framework supports an Agent Directory that:
- Stores OASF-compliant metadata about agents
- Allows agents to discover peers based on capabilities and attributes
This solves the “agent discovery” problem in distributed systems by creating a shared registry of interoperable agents.
Cross-Platform Compatibility – By enforcing schema validation:
- Ensures agents adhere to predefined interaction patterns
- Prevents data format mismatches during communication
This allows agents built on different frameworks (LangChain, AutoGen, etc.) to collaborate seamlessly.
Security Through Standardization – OASF enhances trust in multi-agent interactions by:
- Defining identity verification requirements
- Establishing data integrity checks
This creates a foundation for secure inter-agent communication without proprietary security implementations.
Workflow Coordination – The framework enables complex multi-agent workflows through:
- Agent Manifest standardization: Describes dependencies and deployment requirements
- Workflow Server integration: Executes OASF-defined workflows across diverse agents
This allows orchestration of agents from different providers in coordinated processes.
Ecosystem Scalability – OASF’s attribute-based taxonomy system:
- Organizes agent capabilities hierarchically
- Supports dynamic schema extensions
This accommodates new agent types and use cases without breaking existing integrations.
By solving critical interoperability challenges like data silos, inconsistent APIs, and discovery complexity, OASF reduces integration costs by an estimated 40-60% compared to custom implementations. Its adoption in frameworks like AGNTCY positions it as a foundational standard for the emerging “Internet of Agents” ecosystem.
Conclusion
These open-source standards are designed to address challenges in AI agent interoperability, facilitate easier development and deployment of agentic systems, and create a more connected ecosystem for AI agents. As the field continues to evolve, these standards will likely play a crucial role in shaping the future of AI agent development and collaboration.
