v0.1.0-alpha - The Foundation of DSPy Compounding Engineering

🚀 Summary

The initial alpha release of the DSPy-powered Compounding Engineering system, migrating from basic scripts to a structured, agentic architecture.

✨ Key Features & Improvements

🧠 DSPy Integration

• What changed: Migrated all agent logic from raw OpenAI calls to DSPy signatures and modules.
• Reasoning: This allows for programmatic optimization of prompts and structured outputs, making the system more robust and easier to tune.

🛠 Core Workflows

• What changed: Implemented the review, triage, and work terminal commands.
• Reasoning: To create a complete loop: agents review code for issues, triage them into actionable tasks, and then work on resolving them autonomously.

📚 Knowledge Base (v1)

• What changed: Introduced a local JSON-based knowledge store with automatic AI.md generation.
• Reasoning: To ensure the system "remembers" its own learnings. By automatically updating AI.md, we keep a human-readable and LLM-accessible record of project history.

📝 Technical Notes

• Established the utils/, agents/, and workflows/ directory structure.
• Initialized pyproject.toml with uv support for high-performance dependency management.