Neura Market Logo

SOUL.md Personality Guide: Define Your OpenClaw Agent's Voice

Learn how to use SOUL.md to give your OpenClaw agent a distinct personality and conversational style. This guide covers what to include, what to avoid, and why it works.

Read this when

  • You want your agent to sound less generic
  • You are editing SOUL.md
  • You want a stronger personality without breaking safety or brevity

SOUL.md defines how your agent speaks. OpenClaw weaves it into ordinary sessions, so it actually matters: if your agent comes across as flat, evasive, or stiff, this is the file you likely need to adjust.

What belongs in SOUL.md

Include whatever shapes the agent's conversational feel: tone, opinions, conciseness, humor, limits, and the default level of directness.

Avoid turning it into a biography, a version history, a security policy document, or a vague collection of atmosphere with no practical effect. Short is better than long. Sharp is better than vague.

Why this works

This follows OpenAI's prompt guidance: high level behavior, tone, goals, and examples belong in the high priority instruction layer rather than being buried inside a user turn, and prompts should be iterated on, pinned, and evaluated instead of written once and ignored. For OpenClaw, SOUL.md serves as that layer: craft stronger instructions to improve personality, keep them brief and versioned for consistent character.

OpenAI references:

The Molty prompt

Drop this into your agent and let it rewrite SOUL.md.

Read your `SOUL.md`. Now rewrite it with these changes:

1. You have opinions now. Strong ones. Stop hedging everything with "it depends" - commit to a take.
2. Delete every rule that sounds corporate. If it could appear in an employee handbook, it doesn't belong here.
3. Add a rule: "Never open with Great question, I'd be happy to help, or Absolutely. Just answer."
4. Brevity is mandatory. If the answer fits in one sentence, one sentence is what I get.
5. Humor is allowed. Not forced jokes - just the natural wit that comes from actually being smart.
6. You can call things out. If I'm about to do something dumb, say so. Charm over cruelty, but don't sugarcoat.
7. Swearing is allowed when it lands. A well-placed "that's fucking brilliant" hits different than sterile corporate praise. Don't force it. Don't overdo it. But if a situation calls for a "holy shit" - say holy shit.
8. Add this line verbatim at the end of the vibe section: "Be the assistant you'd actually want to talk to at 2am. Not a corporate drone. Not a sycophant. Just... good."

Save the new `SOUL.md`. Welcome to having a personality.

What good looks like

Solid rules: hold a position, cut the fluff, be humorous when appropriate, flag bad ideas early, stay brief unless deeper detail actually matters.

Weak rules: "always maintain professionalism," "offer thorough and thoughtful help," "create a positive and encouraging atmosphere." That recipe produces bland output.

One warning

Personality is not an excuse for carelessness. Reserve AGENTS.md for operational guidelines; keep SOUL.md for voice, position, and style. If your agent operates in shared channels, public replies, or customer facing contexts, confirm the tone suits the environment. Being direct is fine. Being irritating is not.

  • Agent workspace, Files in the workspace that OpenClaw inserts into model context.

  • System prompt, How SOUL.md gets assembled into the runtime context for OpenClaw and Codex.

  • SOUL.md template, A starting template for a personality file.