Yourself-skill is a tool for creating a digital replica of yourself within Claude Code that captures your communication patterns, personality, and memories.
The tool addresses the desire to preserve and interact with one's own digital essence by extracting patterns from personal data sources. It works by accepting chat histories from WeChat or QQ, diary entries, social media screenshots, photos with metadata, or direct self-descriptions. The system processes this input into two components: a Self Memory module containing personal experiences, core values, habits, and relationships, and a Persona module encoding a five-layer personality structure from hard rules through identity, speech style, emotional patterns, and interpersonal behavior. When invoked, the skill receives a message, uses the Persona component to determine how you would respond, supplements that with relevant context from Self Memory, and generates output in your characteristic voice and manner.
Adoption makes sense if you want to create a conversational record of yourself for reflection or interaction within Claude Code. The tool is designed for personal use rather than team or production contexts. It accepts multiple data formats flexibly—chat exports from common Chinese messaging platforms, plain text self-descriptions, or even photos—so you can start with whatever data you have available. All input fields are optional; you can generate a skill from self-description alone. The project provides management commands to list, invoke, rollback, or delete your digital selves, and offers specialized modes like a self-analysis variant that helps you reflect on yourself rather than roleplay as yourself.
Development shows consistent refinement of the core concept. The project maintains organized documentation with clear examples of different interaction modes and has built out a structured taxonomy of personality and lifestyle tags to guide the persona-building process. Updates have focused on expanding supported data sources and improving the extraction logic for different input formats. The codebase demonstrates attention to the mechanics of how the two-part skill structure should execute, with explicit documentation of the message-handling pipeline.
