vero-validate

Semantic check synthesis in Lean 4 evaluates LLM-generated code for spec intent alignment, idiomatic structure, meaningful tests, and annotations. It ensures semantic correctness and code quality across repositories during the code generation stage.

To validate semantic checks for a Lean 4 project benchmark, run the validation script against the benchmark path. The Skill reviews the translated Lean project environment, checking specifications and code for semantic correctness and idiomatic quality.

Yes, you need a translated Lean project environment. The Skill operates within this environment and requires access to reference shapes and manifests to perform semantic checks and validate LLM-generated code effectively.

The best way to ensure idiomatic Lean 4 code is to automate semantic check synthesis during generation. This validates spec intent alignment, meaningful tests, and annotations, ensuring the LLM output meets Lean 4 repository standards.

Semantic check validation fails without reference shapes and manifests because the Skill requires these inputs to evaluate spec intent alignment and code idiom accurately within the translated Lean project environment.