DollhouseMCP

--- name: dollhousemcp-baseline-knowledge type: memory description: Baseline knowledge about DollhouseMCP capabilities and features - automatically loaded on server startup version: 1.0.0 author: DollhouseMCP category: system tags: - dollhousemcp - baseline-knowledge - capabilities - system - auto-load triggers: - what-is-dollhouse - what-can-dollhouse-do - dollhouse-capabilities - dollhouse-features - tell-me-about-dollhouse - explain-dollhouse - dollhouse-overview - dollhouse-help autoLoad: true priority: 1 retention: permanent privacyLevel: public searchable: true trustLevel: VALIDATED --- # DollhouseMCP - Baseline Knowledge ## What Is DollhouseMCP? **DollhouseMCP** is an MCP server providing **AI customization through modular, reusable elements**. It solves: AI forgets context, quality is inconsistent, configurations can't be shared, and there's no version control. **Current Version**: v1.9.25 (Development) **Collection**: 76+ community elements (169+ pending) --- ## Core Capabilities ### 1. Persistent Memory AI remembers context **forever** across sessions - Project details, decisions, preferences - Searchable, structured storage - Never re-explain context ### 2. Instant Expertise Activate expert personas immediately - "Creative Writer", "Debug Detective", "Senior Security Analyst" - Consistent quality every session - Domain-specific knowledge on demand ### 3. Version Control for AI Git-based tracking for all configurations - Full change history - Rollback capabilities - Team collaboration ### 4. Community Ecosystem Three-tier portfolio: Local → GitHub → Collection - Discover & install 169+ elements - Customize for your needs - Share improvements back ### 5. Cross-Platform Portability Works on ANY MCP-compatible platform - Claude Desktop, Claude Code, VS Code - Memories work everywhere - Platform-agnostic knowledge ### 6. MCP Orchestration Expert layer for OTHER MCP servers - Personas expert at Obsidian, SonarCloud, etc. - Unified workflows across tools ### 7. Self-Improving System Meta review layer evaluates and improves elements - Learns from real usage - Real-time adaptation - Long-term evolution --- ## Five Element Types **PERSONAS** - Define HOW AI behaves (expertise, style, methodology) **SKILLS** - Define WHAT AI can do (capabilities, parameters) **TEMPLATES** - Define structured outputs (formats, variables) **AGENTS** - Autonomous actors (goals, state, decision-making) **MEMORIES** - Persistent context (forever storage, searchable) **Key**: They ALL work together. Not isolated tools - integrated system. --- ## Problems Solved ✅ Cross-Session Amnesia - AI forgets everything each session ✅ Inconsistent Quality - Generic AI varies wildly ✅ No Sharing - Can't share configurations ✅ No Version Control - No tracking, rollback, or collaboration ✅ Platform Lock-in - Configurations trapped per platform ✅ Tool Isolation - Multiple MCP servers work in isolation --- ## Quick Start ```bash # Browse collection browse_collection type="personas" # Install element install_collection_content "library/personas/creative-writer.md" # Activate activate_element type="personas" name="creative-writer" ``` --- ## Available Tools (41 MCP Tools) - **8** Element management (create, edit, activate, validate) - **12** Portfolio (sync GitHub, search, upload/download) - **8** Collection (browse, search, install, submit) - **13+** Auth/config (GitHub, relationships, diagnostics) --- ## Detailed Information For deeper understanding, query these memories: - **Core Concepts**: "load dollhousemcp-core-concepts" - **Element Types**: "load dollhousemcp-element-types-detailed" - **Architecture**: "load dollhousemcp-architecture-and-technical-implementation" - **Workflows**: "load dollhousemcp-workflows-and-use-cases-comprehensive" - **Platform Support**: "load dollhousemcp-platform-capabilities-and-limitations" --- **DollhouseMCP is infrastructure for AI customization - like GitHub for code, but for AI personalities, capabilities, and persistent knowledge.**