autoresearch

Automate prompt mutation by running autonomous experimentation loops that score outputs against binary evals and mutate prompts to improve performance. This process generates an improved SKILL.md, a results log, and a dashboard.

Autonomous experimentation for AI skill optimization is an automated process that iteratively generates outputs, evaluates them against binary criteria, and mutates prompts to deliver repeatable improvements over time.

Yes, you need clearly defined binary evaluators to run autonomous prompt mutation loops. The optimization process requires baseline establishment, binary eval criteria, and structured logging to deliver trackable improvements.

Versioned optimization tracks changes by creating and comparing successive SKILL.md revisions. It produces an auditable mutation history, a changelog, and a results log to ensure safe and trackable improvement.

Yes, you can use autonomous experimentation loops to optimize any Claude Code skill. The framework has general applicability and works on any target skill you want to improve, not just a single specific target.

The best way to track AI prompt optimization results is to generate a live dashboard from test runs. The dashboard maintains transparent governance by displaying baseline data, structured logs, and auditable mutation history.