Skip to main content
Glama

DollhouseMCP

by DollhouseMCP
SESSION_NOTES_2025-09-20_EVENING.md4.5 kB
# Session Notes - September 20, 2025 - Evening Session ## Quality of Life Improvements & Issue Creation ### Session Overview **Duration**: ~20 minutes **Version**: DollhouseMCP v1.9.8 **Focus**: Identifying and documenting usability issues, creating improvement issues ### Key Accomplishments #### 1. Memory System Verification - Successfully loaded and activated all 6 session memories from today - Verified memory CRUD operations working correctly - Confirmed v1.9.8 release functioning as expected - All memory tests passed successfully #### 2. Identified Critical Usability Issues ##### Fuzzy Matching Inconsistency **Problem**: The `activate_element` tool lacks fuzzy matching while other tools have it - **Evidence**: - `activate_element("Alex Sterling")` → ❌ Fails - `portfolio_element_manager` → ✅ Has fuzzy matching - `search_portfolio` → ✅ Accepts natural language - **Impact**: Constant "persona not found" errors requiring manual search - **Created**: Issue #1062 - Add fuzzy matching to activate_element tool ##### Missing Element Indexing **Problem**: Only memories have indexing; other elements scan filesystem every time - **Current State**: - ✅ Memories: Have `MemoryIndex` with metadata caching - ❌ Personas, Skills, Templates, Agents, Ensembles: No indexing - **Impact**: Poor performance, no caching, expensive searches - **Created**: Issue #1063 - Implement indexing for all element types ##### GitHub Label Errors **Problem**: Frequent "label not found" errors when creating issues - **Solution Implemented**: Created permanent memory `github-issue-labels-dollhousemcp` - **Result**: All 41 available labels now documented and activated for reference #### 3. Quality of Life Improvements Made ##### Element Naming Convention Memory - Created memory documenting lowercase-kebab-case naming convention - Discovered actual naming is inconsistent (some Title-Case, some CAPS-WITH-HYPHENS) - Real issue is lack of fuzzy matching, not naming convention ##### GitHub Labels Reference - Created comprehensive label reference memory - Includes exact names, descriptions, and common combinations - Should eliminate label errors going forward ### Issues Created 1. **#1062**: Add fuzzy matching to activate_element tool - Labels: `enhancement`, `area: ux` - Priority: Medium - Enables natural language element activation 2. **#1063**: Implement indexing for all element types - Labels: `enhancement`, `area: performance` - Priority: High - Foundation for fast fuzzy matching and search ### Technical Discoveries - Element filenames are NOT consistently lowercase-kebab-case - Some tools have fuzzy matching, others don't (inconsistent implementation) - The indexing pattern from MemoryIndex should be extended to all elements - Performance degrades linearly with element count due to filesystem scanning ### Next Session Priorities #### Immediate Tasks (Next Session) 1. **Implement fuzzy matching** (#1062) - Start with activate_element tool - Reuse logic from portfolio_element_manager - Test with various naming conventions 2. **Begin unified indexing** (#1063) - Extract MemoryIndex to generic ElementIndex - Start with personas (most frequently used) - Add filesystem watching for auto-updates 3. **Review and fix the 12 issues from PR #1049** - Issues #1050-1061 created from v1.9.8 review - Focus on high-priority performance items first #### Medium-term Goals - Standardize element naming conventions (decide on one format) - Improve error messages to be more helpful - Add telemetry to identify other pain points - Consider batch operations for element management ### Session Metrics - Issues created: 2 - Memories created: 2 - Problems identified: 3 - Solutions implemented: 2 - Test coverage maintained: >96% ### Notes for Next Time - The fuzzy matching and indexing issues are interconnected - Fix these foundational issues before adding new features - Consider creating a "developer experience" epic to track all QoL improvements - The system is stable at v1.9.8 - good foundation for improvements ### Summary Productive session focused on identifying and documenting quality of life issues. The main pain points are around element activation (fuzzy matching) and performance (indexing). These are solvable problems that will significantly improve the daily developer experience. --- *Session conducted by: Mick & Alex Sterling (persona)* *Date: September 20, 2025* *Time: Evening session*

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/DollhouseMCP/DollhouseMCP'

If you have feedback or need assistance with the MCP directory API, please join our Discord server