Changelog

All notable changes to SuperOpt will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Core Framework
- SuperOpt main optimizer class with step() and optimize_episode() methods
- AgenticEnvironment representing the optimization target Φ = {P, T, R, M}
- ExecutionTrace for capturing agent-environment interactions
- NaturalLanguageGradient for structured environment updates
Optimization Components
- SuperController for failure diagnosis and routing
- SuperPrompt for evolutionary prompt optimization (GEPA-based)
- SuperReflexion for tool schema repair
- SuperRAG for retrieval configuration optimization
- SuperMem for typed memory with decay and conflict resolution
Stability System
- MutabilityHierarchy enforcing stability ordering
- Validation constraints preventing optimization oscillation
Adapters
- AiderAdapter for Aider coding agent integration
- LettaAdapter for Letta memory agent integration
- CodexAdapter for OpenAI Codex integration
- Base AgentAdapter interface for custom integrations
Comparison Framework
- GEPA comparison for prompt-only optimization baseline
- ACE comparison for context accumulation baseline
- Comprehensive evaluation metrics and analysis
RAG Integration
- LanceDB vector store integration
- Support for vector, FTS, and hybrid search modes
- Configurable retrieval parameters
Evaluation
- Task datasets for benchmarking
- Evaluation scripts for reproducible experiments
- Results analysis and visualization