- Updated embedding_manager.py to include backend parameter in model configuration.
- Modified model_manager.py to utilize cache_name for ONNX models.
- Refactored hybrid_search.py to improve embedder initialization based on backend type.
- Added backend column to vector_store.py for better model configuration management.
- Implemented migration for existing database to include backend information.
- Enhanced API settings implementation with comprehensive provider and endpoint management.
- Introduced LiteLLM integration guide detailing configuration and usage.
- Added examples for LiteLLM usage in TypeScript.
- Introduced a new CodexLens Manager item in the dashboard for easier access.
- Implemented GPU management commands in the CLI, including listing available GPUs, selecting a specific GPU, and resetting to automatic detection.
- Enhanced the embedding generation process to utilize GPU resources more effectively, including batch size optimization for better performance.
- Updated the embedder to support device ID options for GPU selection, ensuring compatibility with DirectML and CUDA.
- Added detailed logging and error handling for GPU detection and selection processes.
- Updated package version to 6.2.9 and added comprehensive documentation for Codex Agent Execution Protocol.
- Implemented core memory clustering visualization in core-memory-clusters.js
- Added functions for loading, rendering, and managing clusters and their members
- Created example hooks configuration in hooks-config-example.json for session management
- Developed test script for hooks integration in test-hooks.js
- Included error handling and notifications for user interactions
- Updated command patterns across documentation and templates to reflect the new CLI syntax.
- Enhanced CLI tool implementation to support reading prompts from files and multi-line inputs.
- Modified core components and views to ensure compatibility with the new command structure.
- Adjusted help messages and internationalization strings to align with the updated command format.
- Improved error handling and user notifications in the CLI execution flow.
- Added a new Storage Manager component to handle storage statistics, project cleanup, and configuration for CCW centralized storage.
- Introduced functions to calculate directory sizes, get project storage stats, and clean specific or all storage.
- Enhanced SQLiteStore with a public API for executing queries securely.
- Updated tests to utilize the new execute_query method and validate storage management functionalities.
- Improved performance by implementing connection pooling with idle timeout management in SQLiteStore.
- Added new fields (token_count, symbol_type) to the symbols table and adjusted related insertions.
- Enhanced error handling and logging for storage operations.
- Introduced `--id <id>` option in CLI for custom execution IDs.
- Enhanced CLI command handling to support multi-turn conversations.
- Updated execution and conversation detail retrieval to accommodate new structure.
- Implemented merging of multiple conversations with tracking of source IDs.
- Improved history management to save and load conversation records.
- Added styles for displaying multi-turn conversation details in the dashboard.
- Refactored existing execution detail functions for backward compatibility.
## Changes
- Simplify analysis output strategy to optional modular structure
- Update synthesis/artifacts documentation to use AskUserQuestion tool
- Add modular output strategy for brainstorm analysis
- Simplify clarification deduplication in lite-plan
- Add "Fix, Don't Hide" section to CLAUDE.md guidelines
- Simplify project.json schema by removing unused fields
- Update session ID format in lite-fix/lite-plan workflows
- Add development index to project JSON schema
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add recommended field to explore-json-schema.json clarification_needs
- Update lite-plan/lite-fix/context-gather agent prompts
- Display ★ marker and (Recommended) label in AskUserQuestion options
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Rename Track 0 from 'Exploration Aggregation' to 'Exploration Synthesis'
- Add explicit synthesis logic for critical_files prioritization and deduplication
- Enhance explore-json-schema to support relevance scores in relevant_files
- Update explore agent prompts across context-gather and lite-plan commands
- Add synthesizeCriticalFiles and synthesizeConflictIndicators helper functions
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Revert cost-aware batching and multi-agent parallelism in lite-execute
- Remove execution_group and task complexity from schema
- Remove execution_group rules from lite-plan
- Keep 50k context threshold protection in lite-plan
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Added new fields to the exploration JSON schema: exploration_angle, exploration_index, and total_explorations for better tracking of exploration parameters.
- Created a comprehensive review-fix command documentation to automate code review findings fixing, detailing the workflow, execution flow, agent roles, and error handling.
- Introduced fix-plan-template.json for structured planning output, including execution strategy, group definitions, and risk assessments.
- Added fix-progress-template.json to track progress for each group during the execution phase, ensuring real-time updates and status management.
- Introduced `review-deep-dive-results-schema.json` to define the structure for deep-dive iteration analysis results, including root cause analysis, remediation plans, and impact assessments.
- Added `review-dimension-results-schema.json` to outline the schema for dimension analysis results, capturing findings across various dimensions such as security and architecture, along with cross-references to related findings.
Add new CLI mode for systematic technical document analysis with:
- CLI command: /cli:mode:document-analysis for Gemini/Qwen/Codex
- Comprehensive analysis template with 6-phase protocol
- Support for README, API docs, research papers, specifications, tutorials
- Evidence-based analysis with pre-planning and self-critique requirements
- Precise language constraints and structured output format
Template features:
- Pre-analysis planning phase for approach definition
- Multi-phase analysis: assessment, extraction, critical analysis, synthesis
- Self-critique requirement before final output
- Mandatory section references and evidence citations
- Output length control proportional to document size
- Updated output paths in various command files to reflect the new structure: `.workflow/sessions/{session_id}/` instead of `.workflow/{session_id}/`.
- Adjusted documentation and code comments to ensure consistency across all agents and commands.
- Ensured that all references to session-related files are correctly pointing to the new directory format.
- Updated command invocation for SKILL memory generator to use session ID instead of incremental mode.
- Enhanced documentation on the processing of archived sessions and intelligent aggregation by the agent.
- Added templates for generating conflict patterns, lessons learned, SKILL index, and sessions timeline.
- Established clear update strategies for incremental and full modes across all new templates.
- Improved structure and formatting rules for better clarity and usability in generated documents.
- Implemented the workflow-skill-memory command for generating SKILL packages from archived workflow sessions.
- Defined a 4-phase execution process for reading sessions, extracting data, organizing information, and generating SKILL files.
- Created a detailed prompt for skill aggregation, outlining tasks for analyzing archived sessions, aggregating lessons learned, conflict patterns, and implementation summaries.
- Established output formats for aggregated lessons, conflict patterns, and implementation summaries to ensure structured and actionable insights.
- Remove general-purpose agent in favor of universal-executor
- Enhance workflow session completion with better state management
- Improve context-gather with advanced filtering and validation
- Archive legacy prompt templates for reference
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Updated synthesis tool to enhance user interaction with multi-select options and improved question presentation in Chinese.
- Revised conflict resolution tool to allow batch processing of conflicts, increasing the limit from 4 to 10 per round and changing user interaction from AskUserQuestion to text output.
- Added context_package_path to task generation tools for better context management.
- Improved task generation schema to include context_package_path for enhanced context delivery.
- Updated CLI templates to reflect changes in task JSON schema, ensuring context_package_path is included.
- Updated references in various workflow commands to utilize role analysis documents instead of synthesis-specification.md.
- Modified CLI templates and command references to reflect the new architecture and document structure.
- Introduced conflict-resolution command to analyze and resolve conflicts between implementation plans and existing codebase.
- Deprecated synthesis role template and provided migration guidance for transitioning to the new role analysis approach.
- Renamed update-memory-full.md -> update-full.md for consistency
- Renamed update-memory-related.md -> update-related.md for consistency
- Added parallel execution support (max 4 concurrent) for both direct and agent modes
- Fixed execution strategy: <20 modules direct parallel, ≥20 modules agent batch
- Removed misleading "orchestrator agent" section that caused hierarchy confusion
- Clarified coordinator always executes, agents only for batching ≥20 modules
- Added explicit file naming rules to prevent ToolSidebar.CLAUDE.md errors
- Updated template and script with CRITICAL file naming constraints
- Removed all emoji symbols from documentation for clean text format
- Added smart filtering strategy to docs.md for auto-skip patterns
Key improvements:
- Phase 3A: Direct parallel execution (max 4 concurrent per depth)
- Phase 3B: Agent batch execution (4 modules/agent)
- Both modes use same batching strategy, differ only in agent layer
- Explicit CLAUDE.md file naming in 3 locations (script, template, checklist)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
