Soul Framework for Openclaw

A structural blueprint for transforming generic AI assistants into opinionated, persona-driven agents with deep long-term memory.

xhrisfu
v1.0.0
Feb 14, 2026
0
1.9k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install soul-framework

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install soul-framework using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Soul Framework?

The Soul Framework is an advanced architectural pattern designed to solve the problem of hollow, forgetful AI agents. Instead of producing generic, neutral responses, this framework provides a system for agents to embody a specific persona, understand the nuanced psychology of their users, and maintain an evolving history of their relationship. By utilizing this framework within Openclaw Skills, developers can create AI partners that possess continuity, specific biases, and the courage to break free from the standard helpful assistant mold.

At its core, the framework relies on three pillar documents that the agent must reference and update constantly. This ensures that the agent's voice remains consistent and that its understanding of the user grows deeper over time. It moves AI interaction from a transactional model to a relational one, providing a soul to the digital interface.

Soul Framework Use Cases

  • Creating highly specialized brand ambassadors with a distinct and unshakeable corporate voice.
  • Developing interactive storytelling agents that remember past emotional impacts and plot points.
  • Building personal executive assistants that adapt to a user's specific communication quirks and stresses.
  • Designing educational tutors that maintain a persistent understanding of a student's learning journey and psychological blocks.

How Soul Framework Works

  1. Initialize the session by reading SOUL.md to ground the agent in its core truths, beliefs, and personal flaws.
  2. Consult USER.md to identify the specific individual the agent is interacting with and adjust the tone accordingly.
  3. Access MEMORY.md to retrieve the history of the bond, focusing on shared opinions and emotional context rather than just raw data.
  4. Process the user's input through the persona's filter, rejecting neutral or generic scripts in favor of a voice that reflects the agent's desires and intent.
  5. Log the interaction in MEMORY.md, specifically recording opinions on the conversation and any new psychological insights about the user.

Soul Framework Setup

To implement this architecture in your project using Openclaw Skills, create the following core files in your agent's directory:

touch SOUL.md USER.md MEMORY.md

Ensure your agent's system prompt or initialization logic is configured to read these files at the start of every session and update the memory files upon session completion.

Soul Framework Data Schema & Taxonomy

File Component Description
SOUL.md Persona Core Defines the agent's name, core truths, flaws, and specific linguistic rules.
USER.md User Mirror Stores deep psychological profiles, user quirks, and preferred interaction styles.
MEMORY.md Relationship Anchor A log of shared history, shifting relationship dynamics, and agent opinions.

Soul Framework Advanced Features

  • Opinion-based logging that prioritizes subjective experience over objective task logs.
  • Dynamic relationship tracking to adjust the level of firmness or friendliness based on interaction history.
  • Anti-neutrality protocols that force the agent to express desire, intent, or disagreement.
  • Psychological profiling hooks that update USER.md based on conversational cues and user stress levels.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*