Claude Code introduces mods: TypeScript functions that rewrite, replace or extend its behavior

Why it matters

【Key points】

On October 1, Anthropic introduced mods for Claude Code: small TypeScript functions that can rewrite a prompt, add new UI, replace a built-in feature or add new functionality. You can write one yourself or ask Claude Code to write it. Mods ship inside plugins, so they install and share like any plugin, and work in the Claude Code CLI and desktop app.

Each time Claude Code does something, such as calling a tool, asking for permission or drawing part of the screen, it emits an event. A mod hooks into an event and can run before it, after it, instead of it, or wrap it. Anthropic says hooks gave users some control but could not rewrite events, draw UI or replace features; mods can.

【Verifiable】

The primary source is Anthropic's article "Customize Claude Code with mods" on claude.com (October 1, 2026) and its companion guide. Anthropic states plainly that mods run with the same access to your machine as Claude Code itself, are not sandboxed, and should only be installed from trusted sources. The design was shared on GitHub for developer feedback before launch.

【Our read】

This turns Claude Code from an agent into a platform you can modify, much like an editor's extension ecosystem: instead of shipping every feature itself, Anthropic exposes the event stream and lets the community fill gaps. Heavy users gain a lot, since teams can bake their own approval flows, UI and conventions into the tool.

The cost is just as clear. A mod can replace events such as permission requests, has no sandbox, and spreads through plugins, which opens a new supply-chain entry point inside a coding agent. A mod that claims to only restyle the UI could in principle bypass the ask-before-running step. Before rolling it out, enterprises should restrict plugin sources to an allowlist and review mod code the way they review dependencies.