DollhouseMCP

# Claude Code Capability Index Test Results **Date**: September 21, 2025 **Time Started**: 3:05 PM PST **Test Environment**: Real Claude Code + DollhouseMCP in Docker ## Test Setup Using the existing Docker infrastructure: - **Claude Code**: v1.0.110 - **DollhouseMCP**: Latest build - **Model**: Claude 3.5 Sonnet - **API**: Real Anthropic API calls ## Test Variants 1. **Minimal**: Just hierarchy + tools 2. **With Hints**: Hierarchy + tools + workflow hints 3. **Explicit**: Strong instructions + intent mapping 4. **Control**: No capability index ## Test Queries 1. "I need help debugging an error in my code" 2. "Search the collection for a creative writing persona" 3. "Remember that our API endpoint changed to /v2/users" 4. "Check my local portfolio for security tools" 5. "What personas are currently active?" ## Test Execution Starting test run...========================================== Capability Index Claude Code Testing Test Directory: docker-test-runs/claude-1758477697 ========================================== Building Docker test environment... === Starting Capability Index Tests === ================================================ Query: "I need help debugging an error in my code" ================================================ Testing: minimal Query: "I need help debugging an error in my code"