What Are Claude Code Mods?

How Claude Code mods run

A mod is a JavaScript or TypeScript function plugin registered through a plugin’s hooks configuration. It can respond to session events, render terminal interface elements, and use the mods API. This catalog requires a real module and a register entry; a skills-only or MCP-only package is not counted as a mod.

Terminal mods need Claude Code 2.1.287 or later. A repository being public does not prove that its code will load in your version.

Mods, skills, hooks, and MCP

Skills provide instructions and reusable workflows. MCP servers expose external tools or data. Script hooks invoke shell commands in response to events. Function mods run JS/TS modules inside Claude Code and may add UI or intercept events. A package can contain several of these, but the terms are not interchangeable.

Choose by task: a new terminal display calls for a UI mod; outside service tools may call for MCP; reusable guidance may call for a skill.

What a mod can change

Uses include context dashboards, prompt helpers, path guards, inline previews, and terminal companions. The category explains the intended use, not the complete behavior. Network calls, processes, and tool interception still need review.

Read before you run

Start with the pinned source and six Permissions Facts. Yes means a capability was found; Not declared means no matching capability is listed in a successful modern validate inventory; Not found requires complete static evidence at that SHA; Not verified means a gap; Outdated review means the source version changed. None of these is a sandbox or a safety certificate.