openclaw-default-agent-backstory

What This Skill Does

The openclaw-default-agent-backstory skill serves as the foundational architectural framework for your OpenClaw agent. It is designed to move your agent beyond a generic chatbot by systematically building, bootstrapping, and maintaining a high-fidelity workspace identity. By utilizing a structured interview process, this skill populates essential context files that govern how your agent behaves, interacts, and manages long-term memory. It acts as the "system personality" module, ensuring that your agent respects established behavioral guardrails, maintains a consistent growth arc, and stays aligned with your specific workflow requirements.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/msaum/openclaw-default-agent-backstory

Ensure your workspace has the necessary read/write permissions to allow the agent to generate and modify files in the root directory, as this skill strictly manages core configuration files.

Use Cases

• Workspace Onboarding: When spinning up a brand new development environment and you need to define the agent's persona and memory structures from scratch.
• Agent Refinement: When you notice your agent drifting from your preferred communication style and you need to tighten its behavior via identity-refresh mode.
• Context Recovery: When core files like HEARTBEAT.md or SOUL.md are missing or corrupted, this skill acts as a recovery tool to re-seed essential metadata.
• Long-Term Alignment: Use this skill to periodically update the agent's knowledge of your evolving project requirements and personal professional boundaries.

Example Prompts

1. "I'm starting a fresh project. Run the bootstrap process to set up my agent's persona and core memory files."
2. "My agent has been acting a bit too formal lately. Run an identity-refresh and let's update my behavioral guardrails."
3. "Help me refine my agent's backstory. Ask me five questions to get a better handle on my preferred working style and memory retention policies."

Tips & Limitations

• The skill requires access to references/question-bank.md to function optimally; ensure your reference folder is correctly configured.
• Always review the changes made to IDENTITY.md and MEMORY.md after the process completes to ensure no unintentional overrides occurred.
• The skill is non-destructive by default but will rewrite core context files in bootstrap mode; use caution if you have manually customized those files significantly outside of OpenClaw.
• For best results, allow the agent to finish its five-question interview cycle completely before adjusting specific file values manually.