rlm-analyzer 1.2.0 → 1.3.0

package/CHANGELOG.md

@@ -0,0 +1,106 @@
+# 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).
+
+## [1.3.0] - 2026-01-19
+
+### Added
+
+#### Advanced Token Optimization (RLM Paper Implementation)
+- **Context Compression**: Automatically compresses sub-LLM results, achieving 50-70% token savings
+- **Sliding Window History**: Keeps recent turns in full detail while compressing older context
+- **Memory Bank**: Extracts and stores key findings with importance scoring for later synthesis
+- **Adaptive Compression**: Dynamically adjusts compression level based on context usage:
+  - `none` (< 50% usage)
+  - `normal` (50-70% usage)
+  - `aggressive` (70-85% usage)
+  - `emergency` (> 85% usage)
+
+#### New Advanced Features
+- **ParallelExecutor**: Run multiple sub-LLM queries concurrently with retry logic
+- **AdaptiveCompressor**: Context-aware compression that responds to token pressure
+- **ContextRotDetector**: Detects when model loses track of context and injects memory reminders
+- **SelectiveAttention**: Query-based memory filtering for relevance scoring
+- **IterativeRefiner**: Multi-pass analysis for quality improvement (opt-in)
+
+#### Orchestrator Enhancements
+- New feature flags: `enableContextCompression`, `enableParallelExecution`, `enableContextRotDetection`, `enableIterativeRefinement`
+- New accessor methods: `getCompressionMetrics()`, `getContextRotStats()`, `getRefinementHistory()`
+- `executeParallelQueries()` method for batch sub-LLM execution
+
+#### New Exports
+- `ContextManager` and related types
+- All advanced feature classes and their configuration types
+- Subpath exports for `./advanced-features` and `./context-manager`
+
+### Changed
+- Verbose mode now shows compression savings and history optimization
+- Sub-LLM results are automatically compressed before being added to context
+- History is optimized using sliding window when message count exceeds threshold
+
+### Fixed
+- Improved handling of large codebases with many files
+- Better error recovery in sub-LLM execution
+
+---
+
+## [1.2.0] - 2026-01-18
+
+### Added
+- MCP (Model Context Protocol) server for Claude Code and Cursor integration
+- New MCP tools: `rlm_analyze`, `rlm_summarize`, `rlm_architecture`, `rlm_security`, `rlm_dependencies`, `rlm_refactor`, `rlm_ask`, `rlm_config`
+- Binary aliases: `rlm-mcp`, `rlm-analyzer-mcp`
+
+### Changed
+- Improved system prompts for better analysis quality
+- Enhanced error messages for API issues
+
+---
+
+## [1.1.0] - 2026-01-18
+
+### Added
+- **Model Configuration System**: Flexible model selection with priority chain
+  - CLI `--model` flag (highest priority)
+  - Environment variables: `RLM_DEFAULT_MODEL`, `RLM_FALLBACK_MODEL`
+  - Config file: `~/.rlm-analyzer/config.json`
+  - Built-in defaults
+- **Model Aliases**: `fast`, `smart`, `pro`, `flash`, `flash-2`, `flash-2.5`
+- Factory functions: `createAnalyzer()`, `createOrchestrator()`, `getModelConfig()`
+- `rlm test` command to verify API connection
+
+### Changed
+- Default model updated to `gemini-3-flash-preview`
+- Improved CLI help with model configuration details
+
+---
+
+## [1.0.0] - 2026-01-18
+
+### Added
+- Initial release
+- Core RLM orchestrator with REPL execution
+- Sub-LLM delegation via `llm_query()` function
+- Analysis types: summary, architecture, dependencies, security, performance, refactor
+- CLI commands: `summary`, `arch`, `deps`, `security`, `perf`, `refactor`, `find`, `explain`, `ask`, `config`
+- Sandboxed code execution with security patterns
+- Support for 20+ programming languages
+- JSON output mode for scripting
+- Verbose mode for debugging
+
+### Security
+- Blocked dangerous patterns: `eval`, `exec`, `process.`, network calls
+- Read-only analysis - no file modifications
+- API keys stored securely, never logged
+
+---
+
+## [0.1.0] - 2026-01-18
+
+### Added
+- Initial prototype
+- Basic file loading and analysis
+- Gemini API integration

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 BzLabs
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.