An automated framework for holistic evaluation of LLM-generated code using static analysis and sandboxed execution.
-
Updated
May 29, 2026 - Python
An automated framework for holistic evaluation of LLM-generated code using static analysis and sandboxed execution.
Python code analysis engine built on Mathematical Contrapuntalism — security + quality dissonance detection using generating functions, Banach convergence, and path-sensitive taint analysis.
Reproducible static analysis for Python: 8 engines, one schema, provenance attached to every measurement.
To associate your repository with the halstead topic, visit your repo's landing page and select "manage topics."