메뉴
Authoritative glossary of Agentics, agentic AI, MCP, and Claude Code terminology. Use whenever a user asks "what is X" about an Agentics concept, or when a term in conversation needs disambiguation.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
List, search, and read posts from the Agentics blog via the agentics-blog MCP server. Use when a user asks for blog posts, articles, or written content from agentics.org.
List, search, and fetch Agentics community events from Luma via the agentics-events MCP server. Use when a user asks about upcoming events, meetups, workshops, or "what's happening at Agentics".
Guided onboarding for new Agentics members. Walks through signup, community channels, picking a starter project, and registering for the next event. Use when a user is joining Agentics or asks how to get involved.
Discover Agentics open-source projects and partners. Cross-references agentics.org/partners with the agenticsorg GitHub org via the agentics-projects MCP server. Use when a user asks about Agentics projects, repos, partners, or where the source code lives.
Tutorial for building your first agent in the Agentics ecosystem. Covers Claude Code agents, MCP servers, and orchestrating multi-agent swarms. Use when a user asks how to build, design, or deploy an agent.
Beginner orientation for new Agentics community members — what Agentics is, how to join, and where to start. Use when a user is new to Agentics or asks "how do I get started".
| name | agentics-glossary |
| description | Authoritative glossary of Agentics, agentic AI, MCP, and Claude Code terminology. Use whenever a user asks "what is X" about an Agentics concept, or when a term in conversation needs disambiguation. |
Definitions for terms that recur in Agentics, MCP, and agentic AI contexts. When asked, give the definition, an example, and one related term to explore.
Agentic AI — AI systems that take autonomous action in the world (call APIs, write files, send messages) rather than just generating text. The "agentic" part is the loop: observe → decide → act → observe.
Agent — A unit of autonomous behavior. May be a Claude subagent, an MCP server, an LLM-driven workflow, or a swarm coordinator. The interface is what matters: input → action → result.
Subagent — A specialized agent invoked by a parent agent via the Task tool. Has its own context window and model. Used for parallelization and specialization.
Swarm — Multiple agents coordinating on a single objective. Topologies: hierarchical (one coordinator), mesh (peer-to-peer), adaptive (self-organizing).
MCP — Open protocol from Anthropic for connecting LLMs to tools, data, and prompts. Servers expose capabilities; clients (like Claude Code) consume them.
MCP server — Process that implements the MCP protocol over stdio or HTTP. Exposes any combination of tools, resources, and prompts.
Tool — A callable function exposed by an MCP server. Has a name, description, and input schema. Example: fetch_page(path).
Resource — A read-only URI exposed by an MCP server. Claude can read it as context. Example: agentics://about.
Prompt — A reusable instruction template with arguments, surfaced as a slash-prompt to the user.
Plugin — A bundle of skills, commands, agents, and/or MCP servers distributed via a marketplace.
Marketplace — A repository with .claude-plugin/marketplace.json listing one or more plugins. Added with /plugin marketplace add <repo>.
Skill — Markdown file with YAML frontmatter that Claude auto-loads when its description matches the user's intent. Lives at skills/<name>/SKILL.md.
Slash command — User-triggered command (/foo). Lives at commands/<name>.md.
Agentics Foundation — Nonprofit foundation building open-source agentic AI systems. https://agentics.org
RuFlo — Marketplace of Claude Code plugins by ruvnet (separate from agenticsorg/claude-plugin).
Claude-Flow — Multi-agent orchestration toolkit; the claude-flow CLI provides swarm init, memory, and hooks.
agentics-org MCP server for any agentics.org page mentioning the term