Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
A framework for building AI agents with persistent, opinionated personalities. Structures identity through three files: SOUL.md for persona definition, USER.md for user context, and MEMORY.md for relationship history. Agents read these files each session to maintain consistent character across conversations.
Unlike raw system prompts, this framework separates identity, user context, and memory into distinct files that agents actively review, making personality drift and amnesia between sessions less likely.
When to use it
- Building a custom AI character that remembers past interactions
- Making an agent push back on bad ideas instead of just agreeing
- Configuring different agent tones for different user personalities
- Logging emotional and relational context between chat sessions
- Replacing generic assistant behavior with a defined, consistent voice
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: How do I start a new agent session with a defined soul?
- 1Agent reads SOUL.md on startup and internalizes core traits and flaws
- 2Agent reads USER.md to adjust tone based on the specific user's psychology
- 3Agent reviews MEMORY.md to recall previous interactions and relationship dynamics
- 4Agent filters its response through the persona lens before replying
- 5Agent appends new relational observations to MEMORY.md after the session
A response delivered in the agent's defined persona voice, grounded in relationship history rather than generic politeness