Claude Code output styles: change how Claude answers

An output style is a set of instructions that sets Claude's role, tone and response format for a whole session. Instead of asking for "shorter answers" in every prompt, you pick a style once. Claude Code has four built-in styles besides its default, and you can write your own.

The built-in styles

• Default: no style is selected. Claude works from Claude Code's standard instructions for software engineering.
• Proactive: Claude starts work right away and makes reasonable assumptions about routine decisions instead of asking. It also checks with you in the conversation before deleting data or changing a shared or production system. That check is an instruction Claude follows, separate from permission prompts, and your permission mode does not change when you switch.
• Concise: answers lead with the result, with no preamble, narration or recap. Simple questions get one to three sentences. Error reports, failing test output and warnings before destructive actions keep their full content. The documentation says this one needs a recent version.
• Explanatory: Claude does the task as usual and adds short Insight blocks in the conversation that explain why it made its choices. They are not written into your files as comments.
• Learning: the same Insight blocks, and Claude also leaves small pieces for you to write. It marks the spot with a TODO(human) comment in the file, explains what to do, and waits for you.

Switch styles

/output-style concise

Run /output-style with no argument to list the styles and see the current one. You can also run /config and choose Output style. Either way, your choice is saved to .claude/settings.local.json, the local project level. The new style applies from your next message.

You can also set it in a settings file:

{
  "outputStyle": "Explanatory"
}

This value is case-sensitive: Proactive, Concise, Explanatory, Learning. A value that does not match exactly, such as explanatory, silently gives you the Default style (the /output-style command ignores case). To make a style your default everywhere, put outputStyle in ~/.claude/settings.json; a project's own settings override it. See the settings guide.

Write your own

A custom style is a Markdown file with a short header and then your instructions. Save it in ~/.claude/output-styles/ for all your projects, or .claude/output-styles/ for one project. An example:

---
name: Plain English
description: Short answers in plain words, no jargon
keep-coding-instructions: true
---

Explain things in plain words. If a technical term is needed,
say what it means in one short phrase. Lead with the answer
and give more detail only if asked.

All header fields are optional. The file name becomes the style name unless you set name. Set keep-coding-instructions: true if you are only changing how Claude communicates and still want its usual software engineering behavior. Leave it out if Claude will not be doing software engineering at all. A misspelled field is ignored without an error. Claude Code reads style files at startup, so restart it after creating or editing one.

It is an instruction, not a guarantee

A style gives Claude instructions to follow; nothing enforces it. Pick the feature that fits what you want:

• Every answer in one voice, length or format: an output style.
• What Claude should know about your project: CLAUDE.md. See the memory guide.
• Something that must happen every time, such as formatting after each edit: a hook. See the hooks guide.
• Instructions for one kind of task: a skill. A focused helper in its own context: a subagent.

This guide compares CLAUDE.md, skills and hooks in more detail.

Good to know

• A style applies to the main conversation and to a fork of it. Other subagents run their own system prompt, so a style does not change how they answer.
• A style's instructions add input tokens, which prompt caching reduces after the first request. Explanatory and Learning produce longer answers by design, and Concise shorter ones. See the quota guide.
• If a style does not seem to apply, check the spelling and case of the name, and see debugging your configuration.
• Options and version requirements differ between the terminal, the VS Code extension and the desktop app, so check the documentation for yours.

Try it on something harmless

1. Run /output-style to see the list.
2. Run /output-style concise and ask a simple question.
3. Run /output-style default to go back, and compare.

This follows Anthropic's output styles documentation as we read it in October 2026. We are not affiliated with Anthropic. We wrote this from the documentation and have not tested every case. Where this page and the documentation disagree, the documentation wins.

The free loop-stop skill gives long tasks a limit and a stop rule. If you want a second pair of eyes on your CLAUDE.md and settings, the Tune-Up Kit is a checklist for exactly that.