chrisvoncsefalvay

This skill offers a generalized autonomous optimization loop to iteratively improve any measurable artifact, such as code or documents, through structured experiments. It operates as a web runtime package, not assuming subprocess access or unrestricted local files.

This skill provides a generalised autonomous optimisation loop, using soft RLVR to iteratively improve any measurable artifact like code, prompts, or designs. It achieves this by running structured experiments, evaluating results against a multi-dimensional rubric, and learning from each attempt.