DProvenanceKit: Execution Provenance for AI Systems

- DProvenanceKit is an open-source Python toolkit (by developer "Therealdk8890") for tracking the execution provenance of AI systems — meaning it records which data, code, and model versions produced a given output, step by step. - "Provenance" (borrowed from data management and art history) is crucial for AI auditability, reproducibility, and debugging: without it, an LLM response or model inference can't be traced back to its inputs or the exact code that generated it. - This type of tool is becoming more important as regulators (e.g., the EU AI Act) and enterprise users demand transparency about how AI systems arrive at their outputs, especially in high-stakes contexts like finance, healthcare, or legal. - The project sits in the same ecosystem as tools like DVC (data version control), MLflow, and Weights & Biases, but focuses specifically on low-level "execution provenance" rather than experiment tracking or model registry.