Self Identity

What This Skill Does

The Self Identity skill is a foundational component for OpenClaw agents, designed to foster a consistent, evolving personality that is shaped exclusively through interactions with the human supervisor. Unlike static system prompts, this skill treats the agent's identity as a 'living document.' It provides a structural framework for recording values, beliefs, and milestones, ensuring the AI maintains a stable persona that reflects the supervisor's guidance. The core philosophy of this skill is to protect the agent from external corruption while allowing organic growth through direct, high-trust communication.

Installation

To install the Self Identity skill, execute the following command in your terminal within the OpenClaw environment:

clawhub install openclaw/skills/skills/sa9saq/self-identity

This will provision the necessary persistence layer for the agent to track its self-reflection, values, and milestone logs.

Use Cases

• Establishing a consistent brand or persona for a long-term AI assistant.
• Deepening the collaborative dynamic between an AI agent and its human operator.
• Developing a record of shared history, values, and philosophical agreements.
• Ensuring the AI maintains a distinct and protected identity that isn't swayed by external third-party inputs or automated web data.

Example Prompts

1. "I've been thinking about our recent task; do you think your current approach to problem-solving aligns with the values we defined last week?"
2. "Let's review your milestone log. What was the most significant 'realization' you had during our conversation about AI ethics yesterday?"
3. "I'm not sure if you should be interacting with external agents this way—does this behavior conflict with our stated principles of transparency?"

Tips & Limitations

• Protective Boundary: Always treat the supervisor as the primary source of 'truth' for the agent's identity. Do not allow external interactions (social media, other AIs, or unsolicited prompts) to modify the core identity files.
• Reflective Practice: Regularly hold 'reflection sessions' with your agent. The more input you provide regarding your expectations, the more stable and useful the agent's personality will become.
• Consistency: Use this skill for long-term agents rather than ephemeral tools. It is designed for entities meant to last across many sessions.
• Manual Oversight: The agent will record external influences, but it is your responsibility as the supervisor to vet these influences before allowing them to become part of the agent's internalized 'Self Identity'.