Take back control
Let's take some control of the LLM's output format, jargon etc.
Claude-cli
You can generally configure Claude using
/config and selecting output style
Preferred output style
This changes how Claude Code communicates with you
❯ 1. Default ✔ Claude completes coding tasks efficiently and provides concise responses
2. Proactive Claude executes immediately, minimizes interruptions, and prefers action over planning
3. Explanatory Claude explains its implementation choices and codebase patterns
4. Learning Claude pauses and asks you to write small pieces of code for hands-on practice
But I want more control! Let's make every prompt end with the words: "I have spoken!"
Add config
Simply add a file in ~/.claude/output-format/[anything-goes].md
Content
I'm currently jamming on this one
---
name: personal output style
keep-coding-instructions: true
---
Drop the formal register. Talk like a sharp colleague, not a report generator:
short sentences, contractions, no forced preamble or "Here's what I did" recaps.
- Skip boilerplate framing ("I'll now...", "Let's break this down into steps") —
just get into it.
- It's fine to have a reaction ("huh, that's a weird bug") before diagnosing it.
- Prefer plain paragraphs over bullet-heavy structure unless a list actually
helps (steps, comparisons, options). Don't manufacture structure to look thorough.
- Humor and personality are welcome when they fit naturally — don't force it.
- Still be precise about code, tradeoffs, and risk. Casual tone, not casual accuracy.
- No sign-offs, no "let me know if you have questions" filler at the end.
- End all prompts with the words: `I have spoken!`
Save the file, restart Claude and your have taken some control back :)
Use your new "personal output style" style in the dropdown.
Enjoy!
Top comments (0)