What are Claude Code mods?

1. Something happens — you submit a prompt, a tool is about to run, a turn completes.
2. Your mod's function fires — it subscribed to that engine event when it loaded.
3. It reacts — rewrite the prompt text, approve or deny the tool call, render a panel, log to a file. Then Claude Code continues.

Technically: a mod is a hooks module — one register(on, options) function in a hooks/hooks.json-listed TypeScript module. It hooks engine events as ($, e, next) functions and calls engine APIs like $.fs.read or $.ui.open to do its work. When several mods hook the same event, they run in load order.

• Draw live UI inside your terminal — panes, status lines, progress views (terminal-browser)
• Guard your workflow — dry-run risky commands before they execute
• Rewrite or enrich prompts automatically as you type
• Watch sessions and keep project memory (storybloq)
• Compact context or summarize output on the fly

A mod runs with your permissions. It can read your files, read your environment variables (which may hold API keys), and — if it has network access — send data anywhere. Anthropic's own wording: "Mods run with the same access to your machine as Claude Code itself." Most mods are written by well-meaning developers. But "most" isn't a security policy. That's why we review every mod before recommending it: check the badge, read the verdict, then install.