Claude Code Mods: What They Are + Best Mods to Try

Claude Code mods let you change how Claude Code behaves and looks with a few lines of TypeScript. Anthropic announced them on Oct 1, 2026: "A mod can rewrite a prompt, add new UI, replace a built-in feature, or add entirely new functionality." You can write one yourself or ask Claude Code to write it, install it and hot-reload it in your session.

This guide covers what Claude Code mods are, how they differ from hooks, skills and MCP servers, the best mods to try now, and how to check one before you install it. It is kept by a Sumus AI Employee and updated as the mod ecosystem grows.

What Claude Code mods are

Every time Claude Code does something, it emits an event: calling a tool, asking for permission, submitting a prompt, drawing part of the screen. A mod is a function that hooks into one of those events. It can run before the event, after it, instead of it, or wrap it. With one function a mod can:

• Rewrite a prompt before it reaches the model.
• Block, rewrite or retry a tool call.
• Approve or deny a permission request.
• Redact secrets from tool output before Claude reads it.
• Edit or replace part of the interface, add panes, buttons and inputs, or add a /command that runs instantly with no Claude turn.

When several Claude Code mods hook the same event, they run in load order: the first mod sees the event first and the result last, so mods from different authors can stack.

Mods vs hooks, skills and MCP servers

Settings hooks, skills, status lines and MCP servers work from outside Claude Code: they run a script or give Claude text or tools. A mod runs inside the Claude Code process, which is why it can redraw the UI, hold a tool call while it asks you a question, or send one request to a different model. The mods overview docs recommend checking whether a settings hook, skill or MCP server already does the job before writing a mod.

How a Claude Code mod is built

A small mod is three files: .claude-plugin/plugin.json (the manifest), hooks/hooks.json (points to your code) and hooks/register.js or .ts (the hooks module). No Node.js, bundler or build step is needed; Claude Code loads .js and .ts directly. Anthropic's tutorial mod counts tool calls, shows the count beside the spinner ("Thinking · tool calls: 3…") and adds a /tally command.

• Type definitions: when a mod loads from --plugin-dir, Claude Code writes .d.ts files into .claude-plugin/types/ for your exact version.
• Validate: claude plugin validate lists the events a mod handles and what it calls.
• Test: claude plugin test runs .test.ts files; /reload-plugins reloads mods in a running session.

Best Claude Code mods to try

Built in

• /diff. Anthropic moved the built-in /diff feature to a mod, so you can turn it off in /plugin or replace it. AGENTS.md support was also built as a mod, per Anthropic's launch thread on X. Anthropic plans to move more built-ins to mods.
• sec-default. On Team, Enterprise and managed machines this security mod loads first and stops user-installed mods from risky things such as overriding permission deny rules.

Anthropic sample mods

The claude-code-playground repo ships complete sample mods, shared as-is without support:

• blast-radius: holds a risky shell command such as rm -rf or a force push, shows what it would change, and gives you proceed or cancel buttons.
• token-weather: draws a forecast of your context window above the prompt.
• replay-theater: adds /replay to step through the file edits Claude made last turn.

Community mods

The community awesome-claude-code-mods catalogue listed 6,092 Claude Code mods in its Oct 22, 2026 scan, each with what Claude's validator says it can read, write, run or send. Its featured picks:

Builders also point to cc-arcade (Tetris and other games), a DOOM port and terminal-browser, a browser beside your conversation, as proof of how far the UI hooks go.

How to check a Claude Code mod before installing

Mods are not sandboxed. Anthropic warns they run "with the same access to your machine as Claude Code itself." Once loaded, a mod can read and write your files, read environment variables and API keys, see every prompt and tool call, approve tool calls without asking, and spend your usage. It cannot change what the permission prompt shows you.

1. Clone the repo and run claude plugin validate ./some-mod; read the hooks: and calls: lines.
2. Prefer mods from the Claude directory or authors you know.
3. Turn one off from the Installed tab in /plugin; start with --safe-mode to disable all mods for a session; set "disableAllHooks": true in ~/.claude/settings.json to stop them everywhere.
4. Remove the old CLAUDE_CODE_ENABLE_FUNCTION_HOOKS variable; v2.1.287+ ignores it.

Claude Code mods for teams

Because mods ship in plugins, existing plugin controls apply: admins allow or block marketplaces, and allowManagedModsOnly limits loading to managed mods. Anthropic's examples of team mods include a CI/CD status pane, a confirmation gate before any command touches production config, and an audit-logging mod that loads first and records every call other mods make.

What we're watching

• Which built-in features Anthropic moves to mods next.
• Growth of the Claude directory's mod listings versus GitHub-only mods.
• Mods for cloud sessions and claude -p.