A skill teaches an AI assistant how to carry out a repeatable workflow. An MCP server—or, depending on what you mean by “connector,” a connected app—provides access to external tools or services. A plugin packages capabilities such as skills and MCP servers into something people can discover and install. They solve different problems and can be used together.
What is a skill?
A skill is reusable workflow guidance: instructions and optional resources that help an assistant handle a task consistently. OpenAI describes a skill as a folder that can contain a SKILL.md file, reference material, scripts, templates, or other assets. It can specify when to use a workflow, the steps to follow, and what a successful result should include. OpenAI’s skills documentation puts the distinction this way: “A skill provides the workflow around its tools: when to call them, in what order, how to handle incomplete results, and what the final output should contain.”
As an Amazon Associate I earn from qualifying purchases.
A skill does not inherently connect to an external service. It can guide a workflow using tools already available to the assistant, or explain how to use tools supplied by an MCP server. For example, a skill might define a consistent process for preparing a project update; whether it can fetch project data depends on the tools and permissions available separately.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →What is an MCP server or connector?
Model Context Protocol (MCP) servers expose tools and structured capabilities that an assistant can use. Depending on the integration, those capabilities may include service-backed information, authentication, or controlled actions. OpenAI’s documentation summarizes the role as: “An MCP server provides live information and controlled actions.” The server supplies capabilities; it does not necessarily provide the task-specific workflow that tells the assistant how to use them to reach a particular result.
#1 Best Overall
“MCP connector” is not a single, stable product category established across every OpenAI product surface. People may use the phrase to mean an MCP server, a connected app, or another integration that gives an assistant access to external capabilities. It helps to check what the specific product calls the connection and what it actually provides.
How connected apps fit
OpenAI’s Help Center describes an app as a connection between ChatGPT and an external service. As the Help Center explains, “An app connects ChatGPT to an external service, such as Google Drive or Slack, so it can access information or perform supported actions.” An app and an MCP server both concern access to capabilities beyond workflow instructions, but the exact setup and available actions depend on the integration and product surface.
What is a plugin?
A plugin is a packaging and distribution layer, not another name for a skill or an MCP server. OpenAI describes plugins as packages people can discover, install, share, and publish in ChatGPT and Codex. A plugin can contain one or more skills, an MCP server, connected apps, or a combination of these; it may also include optional user-interface elements. OpenAI’s plugin architecture documentation says: “Plugins are the packages people discover, install, share, and publish in ChatGPT and Codex.”
In short, a skill guides a workflow, an MCP server exposes tools, and a plugin can bundle capabilities into an installable experience. A plugin can include a skill without an MCP server, or an MCP server without a skill.
How to choose the right building block
| If you need to… | Use… | Why |
|---|---|---|
| Give the assistant a repeatable procedure, format, or decision process using tools it already has | Skill | It packages workflow instructions and optional supporting resources. |
| Provide service-backed tools, live data, or controlled actions | MCP server or connected app | These provide external capabilities, subject to authorization and access controls. |
| Make a bundle discoverable and installable, or deliver workflow guidance alongside tools | Plugin | It can package skills, MCP servers, and connected apps as an installable experience. |
Start with the smallest setup that meets the need. Add a server or app when the assistant needs external capabilities; add a skill when people need repeatable guidance for using those capabilities; use a plugin when the bundle should be installable and shareable. OpenAI’s plugin architecture guidance likewise recommends starting with the smallest shape and adding capabilities as needed.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How permissions and availability affect the choice
A plugin does not grant access by itself. If a capability depends on an app or service, the user or administrator may need to authorize the connection, and workspace controls or provider requirements may apply. The plugin cannot access content beyond the permissions granted to the user or to an administrator-managed connection. Installation options, directory visibility, and supported surfaces can vary; check the relevant product and workspace settings rather than assuming a capability is available to every account.
Rank #4
Skills may also be managed separately across ChatGPT and Codex. OpenAI’s Help Center documentation on skills in ChatGPT and its plugin quickstart describe product-specific setup, so instructions for one surface should not automatically be assumed to apply to another.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
