vocabulary: "1.0.0" info: provider: Agentic AI Foundation description: Vocabulary for the Agentic AI Foundation (AAIF), covering the MCP standard, goose agent, and open governance for AI agent interoperability. created: '2026-04-19' modified: '2026-04-19' operational: apis: - name: Model Context Protocol (MCP) namespace: mcp version: "1.0" baseUrl: https://modelcontextprotocol.io status: active - name: Goose AI Agent namespace: goose version: "1.0" baseUrl: https://github.com/block/goose status: active resources: - name: tools description: Callable functions exposed by MCP servers to AI agent clients actions: - list - call - describe - name: resources description: Data sources and context objects accessible via MCP actions: - list - read - subscribe - name: prompts description: Reusable prompt templates exposed by MCP servers actions: - list - get - name: agents description: AI agent instances using MCP tools and resources actions: - create - run - extend - name: extensions description: Goose agent extensions providing additional tools and capabilities actions: - list - install - enable - disable actions: - name: list description: Enumerate available items httpMethod: GET pattern: read - name: call description: Invoke a tool with input parameters httpMethod: POST pattern: write - name: read description: Retrieve the content of a resource httpMethod: GET pattern: read - name: describe description: Get the schema and metadata for a tool httpMethod: GET pattern: read - name: install description: Add an extension or tool to an agent httpMethod: POST pattern: write - name: run description: Execute an agent task httpMethod: POST pattern: write schemas: core: - name: MCPTool description: A callable tool definition in the MCP protocol properties: - name - description - inputSchema - annotations - name: MCPResource description: A data resource accessible via MCP properties: - uri - name - description - mimeType - size enums: transport_types: - http - stdio - websocket annotation_hints: - readOnly - destructive - idempotent llm_providers: - anthropic - openai - google - azure - ollama - bedrock capability: workflows: - name: MCP Tool Invocation description: An AI agent discovers and calls an MCP-compatible tool using the standard protocol apis: - mcp personas: - AI Developer domains: - Tool Use - name: Agent Extension Management description: Install and configure extensions in goose to expand its capabilities apis: - goose personas: - AI Developer domains: - Agent Configuration personas: - id: ai-developer name: AI Developer description: Developers building AI agent applications using MCP tools and AAIF standards workflows: - MCP Tool Invocation - Agent Extension Management - id: standards-contributor name: Standards Contributor description: Open-source contributors defining and evolving AAIF protocols and standards workflows: - MCP Tool Invocation domains: - name: Tool Use description: The use of MCP tools to extend AI agent capabilities - name: Agent Configuration description: Installing, configuring, and managing agent extensions and providers crossReference: - resource: tools operations: - list - call workflows: - MCP Tool Invocation personas: - AI Developer - resource: extensions operations: - list - install - enable workflows: - Agent Extension Management personas: - AI Developer