Claude Code Mods Let Developers Rewrite How the Agent Works in TypeScript

Anthropic announced mods for Claude Code on 1 October 2026. A mod is a small TypeScript function that changes how Claude Code works: it can rewrite a prompt, add new interface elements, replace a built-in feature, or add entirely new behaviour. Mods are available today in the Claude Code CLI and desktop app.

The mechanism is event-based. Each time Claude Code performs an action, it emits an event, and a mod can hook in before, after or instead of that action. According to Anthropic, mods can rewrite prompts, block or rewrite tool calls, approve permission requests and redact secrets from output. Several mods can be stacked, running in load order.

You can write a mod yourself, or ask Claude Code to write one for you, and share it as a plugin. Some built-in features, such as the /diff command, now ship as mods that you can replace with your own version.

Anthropic's examples are practical. Teams might add a CI/CD status display, require confirmation before anything touches production, or keep an audit log of what their mods do. For enterprises there is a security mod called sec-default, which gives administrators a starting point for controls.

Our reading is that this moves the tool from configuration to programming. Settings files let you pick options someone else thought of, while mods let a team encode its own rules, such as never pushing to the main branch without a check, or stripping client names from prompts.

That is useful for studios like ours, where different clients have different rules about code and data. It is also a responsibility. A mod that approves permission requests automatically is a powerful thing to install, so treat mods as code to review, not settings to tick.

Anthropic points to a getting started guide and documentation on code.claude.com. We would begin with a read-only mod, such as an audit log, before letting one change prompts or approve anything.