The most important detail in Anthropic’s launch of Mods for Claude Code is this: they run without a sandbox and with your own permissions. Anthropic warns users about it too, and advises installing Mods only from sources they trust.
A Mod is a plugin that changes both how Claude Code looks and how it behaves. Mods act as middleware and execute directly inside the tool.
How far a Mod’s reach extends
Every Mod is a function written in JavaScript or TypeScript that attaches to particular events. Those events cover tool calls, user prompts and UI rendering.
That lets developers place custom panels beside the chat, intercept tool calls or create brand-new commands. For an add-on system, that is a lot of access.
Anthropic relies on the system too. According to the company, several built-in features, such as the /diff command, already run as Mods. When a company builds its own features on a plugin API, the API probably wasn’t bolted on at the last minute.
Managing trust falls to you
Without a sandbox, a Mod can do anything you’re able to do on your machine. Anthropic has published sample Mods on GitHub, so you can read real code before you run anything.
Businesses do get a safeguard. Organizations can decide which Mods are allowed to load, so an IT team can block an unvetted plugin before it ever reaches a developer’s terminal.
Support varies by platform. Mods work in the CLI and the desktop app, but the VS Code extension supports them only in part. If VS Code is your main editor, check what’s available before you build your workflow around a Mod.
The first official Mod keeps an eye on Claude
Anthropic’s first official Mod plugin is called “You Should Know.” It starts a separate agent that monitors Claude's output. When that agent spots important information users might have missed, it sends a “Heads up” message.
The plugin is disabled by default. To try it, enter /plugin enable cc-plugin-you-should-know@builtin. Keep in mind that this adds a second agent that reads everything Claude produces.