Skip to main content
Glama

Agent Knowledge MCP

CHANGELOG.md41.5 kB
# Changelog All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). # Changelog ## [2.2.1] - 2025-08-06 ### 🚀 Enhanced - **Copilot Instructions**: Added mandatory requirement for `create_document_template` tool before `index_document` - **Documentation Templates**: Enhanced workflow guidance for proper document structure validation - **Tool Usage Optimization**: Improved step-by-step workflow for knowledge base documentation - **Release Process**: Successfully merged v2.2.0 enhancements with git workflow optimizations ### 📚 Technical Details - **File**: `.github/copilot-instructions.md` - Enhanced Phase 4 documentation requirements - **Workflow**: Added mandatory `create_document_template` step before document indexing - **Integration**: Seamless integration with existing knowledge base tools - **Git Management**: Proper branch merging and release tagging completed ## [2.2.0] - 2025-08-06 ### ✨ Enhanced - **Tool Parameter Descriptions**: Significantly improved `index_document` tool parameter descriptions to prevent LLM validation errors - **Workflow Guidance**: Added comprehensive workflow reminders and recommendations in tool descriptions - **Force Index Guidance**: Enhanced `force_index` parameter with intelligent usage tips to avoid multiple tool calls - **Document Template Integration**: Added prominent reminders to use `create_document_template` tool first - **AI Enhancement Notes**: Included workflow guidance for better LLM-tool interaction ### 📋 Technical Details - **File**: `src/elasticsearch/sub_servers/elasticsearch_document.py` - **Enhanced Parameters**: Added detailed field specifications for document parameter (Dict[str, Any] → explicit field descriptions) - **Workflow Integration**: Added "💡 RECOMMENDED" and "💡 TIP" guidance throughout tool descriptions - **Testing**: Comprehensive testing confirmed all enhancements work correctly after server restart ### 🔧 Project Standards - **RULE_001**: Created server restart requirement rule for proper testing workflow - **Documentation**: Added `.knowledges/rules/rules.md` with server restart standards - **Version Consistency**: Updated all version numbers across project files ### 🎯 Impact - **Reduced LLM Errors**: Clear parameter descriptions prevent missing field validation errors - **Improved Workflow**: Enhanced guidance leads to better document creation practices - **Better User Experience**: More intuitive tool usage with embedded workflow recommendations ## [2.1.4] - 2025-08-06 ### 🐛 Fixed - **Critical Bug**: Fixed metadata search in `create_index` tool failing to find existing metadata - **Elasticsearch Query**: Removed incorrect `.keyword` suffix from `index_name` field search - **Root Cause**: Field `index_name` is already type `keyword`, so `.keyword` suffix was causing search failures - **Impact**: Index creation now works properly when metadata documentation exists - **Testing**: Verified fix with test index creation after metadata setup ### 📋 Technical Details - **File**: `src/elasticsearch/sub_servers/elasticsearch_index.py` - **Change**: Line 59 - Changed `index_name.keyword` to `index_name` in search query - **Resolution**: Metadata synchronization issue between creation and retrieval resolved ## [2.1.3] - 2025-08-01 ### ✨ Enhanced - **AI Assistant Instructions**: Updated Phase 5 workflow to use `ask_user_advice` tool for next steps guidance - **Publishing Workflow**: Added comprehensive workflow documentation with detailed implementation steps - **Documentation Cleanup**: Removed outdated documentation files for streamlined knowledge base - **Search Requirements**: Enhanced Phase 1 mandatory search requirements with `list_indices` tool enforcement ### 🔧 Fixed - **Configuration**: Removed trailing comma in server version field for proper JSON formatting - **Workflow Instructions**: Clarified mandatory search requirements and tool usage patterns - **Process Improvements**: Better documentation of workflow steps and requirements ### 📚 Documentation - **Workflow Templates**: Enhanced workflow documentation with pseudocode structure - **Cross-References**: Improved linking between different knowledge base components - **Process Guidelines**: Clearer guidance for AI assistant behavior and workflows ## [2.1.2] - 2025-08-01 ### ✨ Enhanced - **Template System**: Updated all knowledge base templates (workflows, rules, memories) with improved structure - **Cross-Reference Support**: Added anchor links and References sections for better knowledge linking - **Tool Naming**: Renamed `ask_mcp_advance` to `ask_mcp_advice` for proper English grammar - **Single File Architecture**: Enforced strict single-file structure for each knowledge type ### 🔧 Technical Improvements - **Templates**: Workflows use pseudocode structure with BEGIN/END, IF/ELSE, WHILE, WHEN - **IDs**: Added unique IDs (RULE_001, MEM_001, WF_001) to all knowledge elements - **Links**: Implemented markdown anchor links for cross-referencing between knowledge types - **Consistency**: All documentation and code updated to use correct tool naming ## [2.1.1] - 2025-07-31 ### 🔧 Fixed - **Version Consistency**: Fixed missing version update in config.default.json from v2.1.0 release - **Release Process**: Ensured all 5 files have consistent version numbers as per release workflow ### 📚 Technical Details - **Complete 5-file update**: pyproject.toml, src/__init__.py, src/config.json, src/config.default.json, CHANGELOG.md - **Process improvement**: Reinforced proper release workflow adherence - **Quality assurance**: Added version verification before publishing ## [2.1.0] - 2025-07-31 ### 🎯 New Features - **Smart Prompting Assistant**: Thêm prompt `smart_prompting_assistant` với khả năng quản lý workflows, rules, và memories trong thư mục `.knowledges` - **Template System**: Chuyển đổi từ hardcoded prompts sang external markdown templates với placeholder replacement - **External Template Files**: Tạo hệ thống template files trong `src/prompts/templates/` với 3 templates: - `workflow_assistant.md` - Quản lý quy trình công việc - `rules_assistant.md` - Quản lý tiêu chuẩn và quy tắc - `memories_assistant.md` - Quản lý thông tin và quyết định quan trọng ### 🔧 Improvements - **Code Simplification**: Giảm complexity của `smart_prompting_assistant()` từ 150+ dòng xuống ~20 dòng - **Maintainability**: Template content có thể được chỉnh sửa mà không cần touch Python code - **User Experience**: Non-programmers có thể dễ dàng chỉnh sửa prompt templates - **Dynamic Loading**: Hệ thống load template động với placeholder `{{user_request}}` replacement - **Error Handling**: Cải thiện error handling cho template loading và validation ### 📚 Documentation & Workflows - **Development Workflow**: Tạo workflow mới `development-without-restart.md` với hot-reloading setup - **Language Preference**: Ghi nhớ preference tiếng Việt của user trong knowledge base - **Template Documentation**: Đầy đủ documentation cho template system và usage ### 🐛 Bug Fixes - **Syntax Errors**: Sửa lỗi emoji Unicode characters trong Python strings - **File Compilation**: Đảm bảo tất cả Python files compile thành công - **Import Issues**: Sửa lỗi import và module structure ### ⚡ Technical Improvements - **Separation of Concerns**: Tách biệt content và logic trong prompt system - **Extensibility**: Dễ dàng thêm template types mới - **File Organization**: Cải thiện cấu trúc thư mục với dedicated templates directory - **Testing**: Thêm validation cho template files và placeholder functionality ## [2.0.7] - 2025-07-30 ### ✅ Verified - **Elasticsearch Snapshot Tools**: Comprehensive testing of all 3 snapshot tools (create_snapshot, restore_snapshot, list_snapshots) - **Repository Management**: Confirmed automatic repository creation and configuration - **Backup & Restore**: Verified selective and full backup operations with proper conflict detection - **Error Handling**: Comprehensive testing of error scenarios and user guidance - **Integration**: Confirmed seamless FastMCP sub-server architecture ### 🔧 Improved - **Tool Quality**: All snapshot tools provide rich formatting and detailed feedback - **User Experience**: Enhanced error messages with actionable troubleshooting steps - **Documentation**: Complete testing results documented in knowledge base - **Reliability**: Enterprise-grade backup and recovery capabilities confirmed ### 📚 Technical Details - **Repository**: backup_repository with filesystem storage - **Snapshots Created**: test_snapshot_2025_07_30 (10 shards), comprehensive_backup_2025_07_30 (47 shards) - **Performance**: Fast operations (0.20-0.40 seconds for backup creation) - **Configuration**: Proper path.repo settings and Docker container integration ## [2.0.6] - 2025-07-26 ### 🚀 Enhanced - **AI-Powered Content Similarity Detection**: Implemented advanced AI-powered content similarity detection using LLM sampling for intelligent duplicate prevention - **Smart Duplicate Management**: Enhanced index_document tool with use_ai_similarity parameter for automatic content analysis - **Intelligent Recommendations**: AI provides UPDATE, DELETE, CREATE, or MERGE recommendations with confidence scores and detailed reasoning - **Enhanced Helper Functions**: Added check_content_similarity_with_ai() function for intelligent content analysis and recommendations ### ✨ Added - **AI Analysis Integration**: Direct integration of AI content analysis into index_document workflow - **Confidence-Based Processing**: High-confidence (>80%) AI recommendations proceed automatically - **Smart Content Organization**: AI helps maintain clean, well-organized knowledge base through intelligent duplicate detection - **Comprehensive AI Feedback**: Detailed reasoning and similarity analysis for better decision making ### 🔧 Improved - **Duplicate Prevention**: Goes beyond title matching to analyze actual content similarity - **Knowledge Base Quality**: AI-powered recommendations improve overall knowledge base organization - **User Experience**: Intelligent guidance reduces manual duplicate checking workload - **Decision Support**: AI confidence scores and reasoning help users make informed choices ### 📚 Technical Details - **AI Integration**: Uses FastMCP Context for LLM sampling with Claude-3-opus, Claude-3-sonnet, GPT-4 model preferences - **Helper Function**: check_content_similarity_with_ai() with elasticsearch search + AI analysis - **Hybrid Approach**: Combines Elasticsearch similarity search with AI content evaluation - **Enhanced Tools**: index_document tool now supports intelligent AI-powered duplicate detection ## [2.0.5] - 2025-07-25 ### ✨ Added - **New Prompt**: Added `copilot_instructions` prompt to provide AI assistants direct access to behavioral guidelines - **Enhanced Prompt Server**: Now includes both `mcp_usage_guide` and `copilot_instructions` prompts - **AI Assistant Guidelines**: Complete behavioral protocols and mandatory workflows accessible via MCP ### 🔧 Fixed - **PyPI Entry Point**: Fixed incorrect entry point in published package (was pointing to non-existent `src.fastmcp_server`) - **Import Error**: Resolved `ModuleNotFoundError: No module named 'src.fastmcp_server'` when installing from PyPI - **Package Consistency**: Ensured all entry points correctly reference `src.main_server:cli_main` ### 🚀 Improved - **Build Process**: Enhanced package building with proper validation of entry points - **Version Management**: Updated version consistency across all package files - **Documentation**: Added PyPI installation troubleshooting guidance - **CLI Output**: Updated to show both available prompts clearly ### 📚 Technical Details - **Entry Point**: `agent-knowledge-mcp = "src.main_server:cli_main"` - **New Helper Function**: `_load_copilot_instructions()` for secure file loading - **Error Handling**: Comprehensive error handling for copilot instructions file access - **Affected Installs**: uvx, pip, and other PyPI-based installations - **Fix Status**: Immediately available on PyPI v2.0.4+ ## [2.0.3] - 2025-07-25 ### 🔧 Fixed - **Version Consistency**: Updated pyproject.toml version from 2.0.2 to 2.0.3 for consistency - **Build Process**: Ensured all version references are synchronized across package files - **PyPI Publishing**: Successfully published with consistent version management ### 🚀 Improved - **Package Metadata**: Enhanced package configuration with proper version alignment - **Changelog**: Added comprehensive release documentation with build improvements ### 📚 Technical Details - **pyproject.toml**: Version updated to maintain consistency with src/__init__.py - **Build Output**: Generated agent_knowledge_mcp-2.0.3-py3-none-any.whl and .tar.gz - **PyPI URL**: https://pypi.org/project/agent-knowledge-mcp/2.0.3/ ## [2.0.0] - 2025-07-24 ### 🎉 MAJOR RELEASE: FastMCP Migration This is a major version release with significant architectural changes and breaking changes. ### ⚠️ BREAKING CHANGES - **Complete Migration to FastMCP Framework**: Replaced legacy MCP implementation with modern FastMCP - **New Server Architecture**: Redesigned server structure with `main_server.py` as primary entry point - **Confirmation Middleware**: Implemented FastMCP-native confirmation middleware with better user experience - **Enhanced Prompt Server**: Added comprehensive MCP usage guidance through integrated prompt server - **Configuration Changes**: Updated configuration structure for FastMCP compatibility ### Added - **FastMCP Framework Integration**: Complete migration from legacy MCP to FastMCP - Modern middleware support with separation of concerns - Enhanced error handling and user interaction patterns - Better tool registration and management - Improved performance and reliability - **New Main Server Architecture**: `src/main_server.py` - Centralized server composition - FastMCP middleware integration - Streamlined tool mounting and configuration - Better separation between admin and core functionality - **Advanced Confirmation Middleware**: `src/middleware/confirmation_middleware.py` - FastMCP-native middleware implementation - Configuration-driven confirmation rules - Simplified user experience with clean confirmation dialogs - Python 3.13 compatibility with proper typing - **Comprehensive MCP Usage Guide**: `src/resources/mcp_usage_instructions.md` - 15,000+ character comprehensive usage documentation - 4 real-world workflow scenarios - 3 progressive difficulty tutorials - Complete tool reference with examples - Troubleshooting and best practices sections - **Enhanced Prompt Server**: `src/prompts/prompt_server.py` - MCP usage guidance delivery through LLM interactions - English language support for professional documentation - Integration with comprehensive usage instructions - **Ask User Advice Tool**: Enhanced human-in-the-loop decision making - Multi-level urgency support (low/normal/high/urgent) - Comprehensive request formatting - Context-aware user guidance requests ### Changed - **Server Entry Point**: Primary server now runs through `main_server.py` instead of legacy server - **Middleware Architecture**: Moved from inline confirmation to proper FastMCP middleware patterns - **Tool Organization**: Better separation between admin tools and core functionality - **Configuration Management**: Enhanced configuration loading and validation - **Error Handling**: Improved error messages and user feedback throughout the system ### Enhanced - **User Experience**: Simplified confirmation dialogs and clearer user communication - **Documentation**: Comprehensive usage guides and step-by-step tutorials - **Development Experience**: Better code organization and maintainability - **Performance**: FastMCP framework provides better performance and resource management - **Reliability**: Enhanced error handling and graceful failure recovery ### Migration Notes - **Existing Users**: This version includes breaking changes in server architecture - **Configuration**: May require configuration updates for new FastMCP structure - **Tool Usage**: All existing tools remain functional with improved user experience - **Installation**: Use `uvx upgrade agent-knowledge-mcp` or reinstall for clean migration ### Technical Details - **Framework**: FastMCP 0.15.0+ with full middleware support - **Python**: 3.10+ with enhanced 3.13 compatibility - **Architecture**: Modular design with proper separation of concerns - **Middleware**: Confirmation, prompt delivery, and extensible architecture - **Documentation**: Professional English documentation with comprehensive examples ## [1.0.28] - 2025-07-13 ### Security - **CRITICAL SECURITY FIX**: Fixed user_response tool auto-calling vulnerability - Agent was bypassing confirmation system by auto-calling user_response without user permission - Updated user_response tool description with explicit warning: "⚠️ CRITICAL: This tool MUST ONLY be called by the USER, NEVER by the agent!" - Enhanced parameter descriptions to prevent automatic response setting - Verification confirms agents now properly ask user for approval instead of auto-calling - Complete resolution of confirmation system bypass vulnerability ### Changed - Enhanced security constraints in MCP tool definitions - Improved agent behavior compliance with permission protocols ## [1.0.27] - 2025-07-13 ### Added - **Enhanced Confirmation System Configuration**: Improved configuration management for user confirmation requirements - Fixed duplicate configuration structure in confirmation section - Enabled confirmation for `index_document` operations by default for enhanced security - Streamlined configuration structure with proper validation - **Configuration System Restoration**: Complete revert to original working state after user feedback - Restored `self.default_config` with hard-coded values in ConfirmationManager - Maintained both config.json and config.default.json with identical confirmation sections - Ensured proper backup/restore functionality for server upgrades - **Test File Organization**: Improved project structure organization - Moved test files to appropriate tests directory for better organization - Enhanced project structure following Python best practices ### Fixed - **Configuration Duplication**: Removed redundant configuration entries - **File Organization**: Proper placement of test files in tests directory - **Configuration Validation**: Fixed malformed JSON structure in confirmation configuration ### Changed - **Default Security Posture**: Index operations now require confirmation by default - **File Structure**: Test files properly organized in tests directory - **Configuration Management**: Streamlined confirmation system configuration ## [1.0.26] - 2025-07-12 ### Added - **Time-Based Search Enhancement**: Comprehensive time filtering capabilities for search operations - `date_from` parameter: Filter documents from specific date (YYYY-MM-DD, YYYY-MM-DD HH:MM:SS, or relative: '7d', '1w', '1m') - `date_to` parameter: Filter documents until specific date (YYYY-MM-DD, YYYY-MM-DD HH:MM:SS, or 'now') - `time_period` parameter: Quick time shortcuts ('today', 'yesterday', 'week', 'month', 'year') - `sort_by_time` parameter: Control time-based sorting ('desc' or 'asc') - **Relative Date Support**: Natural language date parsing for flexible time filtering - '7d', '1w', '1m' style relative dates - Automatic conversion to Elasticsearch-compatible date ranges - Support for 'now' as end date reference - **Intelligent Time-Based Sorting**: Dynamic sorting behavior based on time filtering usage - Time filtering enabled: Sort by time first, then relevance - No time filtering: Sort by relevance first, then recency (existing behavior) - **Enhanced Search Guidance**: Time-specific suggestions when no results found with time filters ### Enhanced - **Search Tool Schema**: Extended with comprehensive time-based parameters and documentation - **Search Result Descriptions**: Clear indication of applied time filters and sorting behavior - **Error Handling**: Time-specific suggestions for broadening search when time filters yield no results - **User Experience**: Contextual guidance for time filter usage and optimization ### Technical - **Date Parsing Engine**: Robust parsing for multiple date formats and relative date expressions - **Elasticsearch Integration**: Proper range filters and bool queries for time-based search - **Backward Compatibility**: All existing search functionality preserved, time parameters are optional ## [1.0.25] - 2025-07-12 ### Added - **Enhanced Error Pattern Matching**: Comprehensive detection of Elasticsearch index errors including `index_not_found_exception` and `no such index` patterns - **4-Step Agent Guidance**: Structured guidance for agents when encountering index-related errors with specific tool recommendations - **Improved Error Messages**: Clear distinction between index errors and document errors with actionable next steps ### Enhanced - **Search Error Handling**: Now provides specific 4-step guidance when searching non-existent indices - **Agent Experience**: From generic "Unknown Error" to detailed, actionable error messages with tool suggestions - **Pattern Recognition**: Supports multiple error message variations including "does not exist" formats ### Fixed - **Error Handler Coverage**: All Elasticsearch handlers now have comprehensive pattern matching for better error detection - **User Experience**: Search operations now provide helpful guidance instead of cryptic error messages ## [1.0.24] - 2025-07-12 ### Added - **Knowledge Base Reorganization Guidance**: Automatic detection when search returns too many results (15+ documents) - **Intelligent Search Result Analysis**: Extracts topics, content types, and priority distribution from search results - **User Collaboration Workflow**: Step-by-step guidance for knowledge base reorganization with user confirmation - **Duplicate Prevention Enhancement**: More explicit guidance in all index document functions to prevent duplicate creation ### Enhanced - **Search Function Intelligence**: Now provides specific reorganization suggestions and user collaboration templates - **Agent Guidance Clarity**: Enhanced visual prominence with warning levels (CRITICAL, MANDATORY, IMPORTANT) - **Knowledge Quality Management**: Promotes "quality over quantity" approach with targeted cleanup recommendations ## [1.0.23] - 2025-07-11 ### Added - **Enhanced Agent Error Guidance**: Search errors now provide 4-step specific guidance for AI agents when indices don't exist - **Content Management Strategy**: Intelligent guidance for when to create files vs index content directly (1000+ character threshold) - **Search Optimization**: No-results scenarios trigger comprehensive keyword expansion suggestions and user collaboration prompts - **Dual Result Sorting**: Search results now sorted by both relevance score and recent update time for better discovery ### Enhanced - **File Permission Handling**: Enhanced error messages with specific tool suggestions and fallback strategies for permission issues - **Agent Workflow Integration**: All error handling now includes actionable next steps specifically designed for AI agent workflows ## [1.0.22] - 2025-07-09 ### Added - **Section-Specific Intelligent Merge**: Enhanced configuration merge algorithm with differentiated handling for different config sections - **reset_config Tool**: New manual reset tool that overwrites config with defaults while creating timestamped backups - **Deprecated Settings Detection**: Automatic filtering of deprecated settings using pattern detection for 'old_', 'deprecated_', 'legacy_' prefixes - **User-Only Settings Preservation**: Intelligent merge now preserves user settings even if they don't exist in new config ### Enhanced - **server_upgrade Tool**: Now automatically backs up and intelligently restores user configuration during upgrades - **Configuration Management**: Two-tier approach with manual reset vs automatic intelligent restore - **LATEST CONFIG Sections**: server, schema, version sections always use newest config for compatibility - **INTELLIGENT MERGE Sections**: security, elasticsearch, logging sections preserve user settings while adding new features ### Removed - **restore_config Tool**: Replaced by automatic restoration in server_upgrade process ### Technical - **Algorithm Enhancement**: Recursive merge with section-aware logic for optimal configuration handling - **Comprehensive Testing**: Added extensive test coverage for section-specific behavior and real-world upgrade scenarios ## [1.0.21] - 2025-07-09 ### Enhanced - **Schema Configuration**: Added `content` field to `required_fields` in default configuration schema - **Configuration Consistency**: Ensured content field is properly included in document schema validation - **Documentation Updates**: Improved schema documentation and examples ### Fixed - **Configuration Schema**: Resolved inconsistency where content field was missing from required_fields list - **Schema Validation**: Enhanced schema validation to properly handle content field requirements ## [1.0.20] - 2025-07-09 ### BREAKING CHANGES ⚠️ - **Strict Configuration Mode**: Removed ALL fallback mechanisms from configuration loading. Server now requires proper `config.json` with `document_schema` and `document_validation` sections or will fail to start with clear error messages. - **No Emergency Fallbacks**: Eliminated `EMERGENCY_FALLBACK_SCHEMA` and all silent configuration defaults. Configuration errors now cause immediate RuntimeError with detailed fix suggestions. ### Added - **Configuration Backup/Restore System**: Implemented comprehensive backup and restore system for server upgrades - `config.default.json` - Automatic backup file created from current configuration - `restore_config` - New admin tool to manually restore configuration from backup - **Smart Fallback**: Automatic fallback to `config.default.json` when `config.json` is missing after server upgrades - **Clear User Messaging**: Explicit notifications when fallback configuration is used ### Enhanced - **Strict Error Handling**: Implemented zero-tolerance configuration validation with comprehensive error messages and solution guidance - **Boolean Type Conversion**: Enhanced validation config loading to properly handle string boolean values (`"false"` → `false`) from JSON configuration - **Production Safety**: Server will fail fast on startup if configuration is invalid, preventing silent degradation in production environments - **Upgrade Resilience**: Server upgrades no longer cause configuration loss with automatic backup/restore system ### Documentation - **Comprehensive Guide**: Added `strict-configuration-validation-implementation-v1020.md` with complete implementation details, testing results, and migration guidance - **Backup/Restore Guide**: Added `configuration-backup-restore-system-v1020.md` with detailed upgrade workflow and recovery procedures - **Configuration Examples**: Updated all config examples (`config.json.example`, `config.example.json`) with proper document_schema sections ### Technical Improvements - **Zero Fallback Architecture**: Complete removal of fallback code paths ensures predictable behavior and prevents configuration drift - **Enhanced Error Messages**: Configuration errors include exact file paths, missing sections, and specific fix instructions - **Runtime Safety**: Strict validation prevents server startup with incomplete or invalid configurations - **Multi-tier Fallback**: Controlled fallback system for server upgrade scenarios while maintaining strict validation in normal operation ## [1.0.19] - 2025-07-08 ### Fixed - **Function Signature Bug**: Fixed critical bug in `get_example_document()` function that was causing "takes 0 positional arguments but 1 was given" error when strict validation was enabled during `index_document` operations - **Document Schema Validation**: Removed duplicate `format_validation_error` functions that were causing conflicts in document validation workflow - **Test Infrastructure**: Updated test import paths for proper module loading without MCP dependencies ### Enhanced - **Error Handling**: Improved error messages for validation failures with proper context handling - **Strict Validation**: Strict schema validation now works correctly without function signature errors - **Knowledge Base**: Added comprehensive documentation of the bug fix for future reference ## [1.0.18] - 2025-07-09 ### Fixed - **Index Document Bug**: Fixed critical bug in `handle_index_document` function where documents were validated but not actually indexed to Elasticsearch when `validate_schema=True`. The function now properly proceeds to index documents after successful validation. ## [1.0.17] - 2025-07-07 ### Changed - **Packaging**: Added `config.json` and example files to the package distribution in `pyproject.toml`. This makes the server more robust by providing a fallback configuration if a local `config.json` is missing or invalid when run via `uvx`. ## [1.0.16] - 2025-07-07 ### Fixed - **Version Consistency**: Corrected the `__version__` in `src/__init__.py` to match the `pyproject.toml` version. The `1.0.15` release had an incorrect internal version number. ## [1.0.15] - 2025-07-07 ### Fixed - **Tool Definition Bug**: Corrected the `inputSchema` for the `update_config` tool in `src/tools.py` to match the actual implementation in `src/admin_handlers.py`. This resolves a critical bug that made the tool unusable. ### Added - **Knowledge Base Documentation**: - Added a detailed guide on how to configure the MCP client for local development using `.vscode/mcp.json`. - Added a comprehensive guide for the Python package release process. ## [1.0.14] - 2025-07-07 ### Added - **New get_prompt_guidance tool**: Provides GitHub link to copilot instructions for effective MCP server usage - Simple English message directing users to comprehensive prompting documentation - Zero-parameter tool for easy access to guidance resources ### Enhanced - Admin tools category expanded with user guidance functionality - Improved user onboarding experience with direct access to effective prompting strategies ## [1.0.13] - 2025-07-07 ### Testing - **Auto-Upgrade Workflow Validation**: Final testing of improved server_upgrade tool - Testing complete PyPI API integration and version-specific installation - Validating fallback mechanisms for UV index sync delays - Comprehensive end-to-end auto-upgrade testing from 1.0.10 → 1.0.13 ### Technical - No functional changes, version bump for testing improved auto-upgrade implementation - All enhanced auto-upgrade functionality operational and ready for production use ## [1.0.12] - 2025-07-07 ### Enhanced - Improved server_upgrade tool with robust fallback mechanism - Direct version-specific installation from PyPI API - Clean cache + force install approach for reliable upgrades - Better error handling and upgrade detection - Fallback to latest available version when specific version fails ### Fixed - Auto-upgrade reliability issues with UV index sync delays - Enhanced upgrade detection from installation output - Better user messaging for upgrade success/failure ## [1.0.11] - 2025-07-07 ### Enhanced - Improved server_upgrade tool with direct cache clean + force install approach - Better upgrade reliability: `uv cache clean && uv tool install agent-knowledge-mcp==version --force` - Enhanced output parsing to detect successful upgrades vs reinstalls - Improved error handling with detailed manual fallback instructions ### Testing - Validated new upgrade approach: cache clean + specific version install - Confirmed reliability over standard `uv tool upgrade` command - Better handling of PyPI propagation delays ## [1.0.10] - 2025-07-07 ### Improved - Enhanced server_upgrade tool with direct upgrade attempt before cache clean - Better error handling and fallback mechanisms for auto-upgrade - Improved user instructions for manual upgrade when needed - Added PyPI propagation delay handling ### Testing - Comprehensive auto-upgrade workflow testing - Validated cache cleaning functionality - Confirmed version detection accuracy ## [1.0.9] - 2025-07-07 ### Testing - **Auto-Upgrade Feature Final Test**: Version bump to test complete auto-upgrade workflow with MCP server tools - Test server_status detecting version 1.0.8 vs latest 1.0.9 - Test server_upgrade tool performing cache clean and instructing user reload - Validate end-to-end auto-upgrade experience using built-in MCP server tools ### Technical - No functional changes, version bump for comprehensive auto-upgrade testing - All auto-upgrade functionality ready for production use ## [1.0.8] - 2025-07-07 ### Fixed - **Server Upgrade Tool**: Fixed detection to use `uv tool list` instead of `uvx list` - Corrected package installation verification for modern uv tool workflow - Updated error messages to reference correct installation method - Enhanced compatibility with current uv tool ecosystem ### Technical - All admin tools now properly support uv tool installation method - Improved error handling and user guidance in server_upgrade tool ## [1.0.7] - 2025-07-07 ### Testing - **Auto-Upgrade Feature Test**: Version bump to test the complete auto-upgrade workflow - Test server_status tool detecting version mismatch - Test server_upgrade tool with cache cleaning and user instructions - Validate complete upgrade cycle: detect → clean → reload → verify ### Technical - No functional changes, version bump for testing auto-upgrade feature - All existing functionality remains unchanged ## [1.0.6] - 2025-07-07 ### Fixed - **Version Detection**: Fixed version detection in uvx environment for server_status tool - Enhanced `importlib.metadata` usage for proper package version detection - Improved import fallback logic for different runtime environments - Fixed installation method detection using correct `uv tool list` command - **Auto-Upgrade Feature**: Version detection now works correctly enabling proper auto-upgrade workflow - Server can now detect current version vs latest PyPI version - Upgrade recommendation and workflow function properly - Clean cache and reload workflow operational ### Technical - Removed deprecated `pkg_resources` dependency - Improved error handling in version detection logic - Enhanced compatibility with uvx installation method ## [1.0.5] - 2025-07-07 ### Removed - **Server Uninstall Tool**: Removed `server_uninstall` tool and handler for improved safety - Removed `handle_server_uninstall` function from admin_handlers.py - Removed `server_uninstall` tool definition from tools.py - Removed import and mapping from server.py - Users should manually uninstall via `uvx uninstall agent-knowledge-mcp` if needed ### Changed - **Cross-Platform Compatibility**: Removed emoji characters from all print statements for better Windows terminal compatibility - Updated print statements across all Python files to use plain text - Improved compatibility with older Windows terminal environments - Maintained functionality while ensuring consistent display across platforms - **Server Upgrade Workflow**: Fixed `server_upgrade` function to use proper uvx workflow - Changed from non-existent `uvx upgrade` to `uv cache clean` approach - Added proper user instructions for restarting Claude Desktop client - Improved error handling and user guidance for update process ### Fixed - Corrected uvx upgrade process since uvx doesn't support direct upgrade commands - Enhanced Windows terminal compatibility by removing emoji dependencies - Improved user experience with clearer update instructions ### Technical - Reduced total tool count from 34 to 33 tools - Maintained all core functionality while improving safety and compatibility - Enhanced code maintainability by removing complex uninstall logic ## [1.0.4] - 2025-07-05 ### Added - **Server Management Tools**: Three new admin tools for MCP server self-management - `server_status` - Check server version, installation method, and available updates with PyPI integration - `server_upgrade` - Upgrade MCP server via uvx with comprehensive error handling and user guidance - `server_uninstall` - Safely uninstall MCP server via uvx with confirmation requirements - **Version Detection**: PyPI API integration for automatic update checking in agent workflows - **Enhanced Admin Capabilities**: Expanded from 8 to 11 admin tools for complete server lifecycle management ### Changed - Updated tool count from 31 to 34 tools total - Enhanced README.md with new admin tool documentation - Improved admin_handlers.py with uvx command integration and subprocess management - Updated server.py with new tool handler registrations ### Fixed - Added proper error handling for uvx availability and installation verification - Enhanced safety mechanisms with confirmation requirements for destructive operations ### Security - All server management operations validate uvx installation method before proceeding - Confirmation required for server uninstall to prevent accidental removal - Comprehensive error messages guide users for proper installation and usage ## [1.0.3] - 2025-07-05 ### Added - **Community Support Features**: Complete sponsorship and contribution system - GitHub Sponsors integration with `.github/FUNDING.yml` - Buy Me Coffee integration (https://coff.ee/itshare4u) - Multiple sponsorship platforms (PayPal, Ko-fi, GitHub Sponsors) - Sponsor tier system with clear benefits and recognition - Comprehensive CONTRIBUTING.md with financial support guidelines - **Enhanced README**: Auto-update installation buttons and sponsorship sections - Auto-update installation buttons for VS Code with `--upgrade` flag - Multiple configuration options (stable, pinned, latest, development) - Detailed update instructions for all AI assistant platforms - Beautiful sponsor badges and call-to-action sections - **Strict Schema Validation**: Configurable strict mode to prevent extra fields in documents - **Enhanced Configuration Management**: New tools for complete config modification and validation - `get_config` - View complete configuration - `update_config` - Modify entire configuration with validation - `validate_config` - Validate configuration before applying - **Document Validation Controls**: Fine-grained control over schema enforcement - `strict_schema_validation` - Enable/disable strict mode - `allow_extra_fields` - Control extra field behavior - `required_fields_only` - Enforce only required fields - `auto_correct_paths` - Automatic path normalization - **Enhanced Document Schema**: Support for different validation levels for knowledge base vs custom documents ### Changed - Updated README.md with comprehensive documentation of new features - Enhanced tool count from 28 to 31 tools - Improved error messages with clear validation feedback - Enhanced `validate_document_structure()` function with `is_knowledge_doc` parameter ### Fixed - Resolved issue where agents could add arbitrary fields bypassing schema validation - Improved configuration loading with better error handling ### Deprecated - `get_allowed_directory` and `set_allowed_directory` are now deprecated in favor of comprehensive config management tools (but still functional for backward compatibility) ### Testing - Added `test_strict_validation.py` for testing strict schema validation - Added `demo_config_management.py` for demonstrating new config features - Moved test files to proper `tests/` directory structure ## [1.0.2] - 2025-01-03 ### Added - Initial PyPI publication - Comprehensive MCP server with 28 tools - Elasticsearch integration with full CRUD operations - File system management with cross-platform support - Version control integration (Git/SVN) - Document validation and schema enforcement - Security controls and sandboxed operations ### Features - 9 Elasticsearch tools for search and document management - 11 File system tools for comprehensive file operations - 5 Administration tools for system management - 3 Version control tools for Git/SVN operations - Complete MCP protocol compliance - Auto-setup capabilities for Elasticsearch ### Documentation - Comprehensive README with installation and usage examples - Test suite with multiple demo workflows - Cross-platform compatibility documentation - Security and privacy guidelines ## [1.0.1] - 2025-01-02 ### Fixed - Initial package structure and dependencies - Cross-platform path handling improvements ## [1.0.0] - 2025-01-01 ### Added - Initial release of AgentKnowledgeMCP - Basic MCP server functionality - Core Elasticsearch operations - File system operations - Document validation framework

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/itshare4u/AgentKnowledgeMCP'

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