Claude Code adds Mods: plugins can now change the interface

On October 1, 2026, Anthropic shipped "Mods" in Claude Code version 2.1.287: plugins can now modify Claude Code's behavior at a deeper level. A mod is made of JavaScript or TypeScript event handlers that run inside Claude Code itself.

That is more than teaching Claude a new skill: you can now add a new interface, command or rule to Claude Code itself.

What can a mod do?

• Draw an interface: add a pane beside the transcript, a band above the prompt, buttons and text fields.
• Step into a tool call: hold a tool before it runs, ask the user, or refuse it.
• Add its own command: a /command that runs immediately, without starting a Claude turn.
• Replace existing features: for example, /diff now ships inside Claude Code as a mod.

Anthropic also added a built-in sample mod called "You should know": while Claude works on a longer task, a side agent watches and shows a note about things you or Claude might miss. It is off by default and is enabled with /plugin enable cc-plugin-you-should-know@builtin.

Where does it run?

Mods need v2.1.287 or later and are on by default. Organizations can limit which mods load through managed settings.

What we think at Vinci Studio

For a small agency, this means we can tune Claude Code to the way we work, for example by showing a risky command before it runs. There is no need to rush, though: the feature is new and the ecosystem is young.

One thing to watch

Anthropic's own documentation says it plainly: a mod is code that runs with your permissions; it can read and write your files, start processes and make network requests, and it is not sandboxed. Install only from authors you trust, and before installing you can list what it does with claude plugin validate.

Sources: Anthropic Claude Code docs (Mods overview and changelog), GitHub anthropics/claude-code release notes, Releasebot, byteiota.