autoresearch

Autonomous code-quality improvement loops work by scanning codebase metrics to identify opportunities, scaffolding workflows, and executing iterative changes. They enforce a contract via program.md, track measurements, and revert non-improving changes to ensure safe autonomous iteration.

To start autonomous iterations, run the /autoresearch command to scan your project, surface improvement loops, and begin autonomous iterations toward measurable code quality gains. This scaffolds workflows and executes iterative changes within a defined scope.

Yes, the autonomous iteration process automatically reverts non-improving code changes using git. It tracks measurements throughout the iterative changes and reverts any edits that fail to improve the defined code-quality metrics.

Yes, autonomous code improvement loops can be applied to software projects with customizable codebases. They execute iterative changes strictly within a defined scope, enforcing a contract via program.md to ensure safe and measurable iterations.

Autonomous iteration tracks code quality metrics by scanning the codebase, executing iterative changes, and logging outcomes. It enforces a contract via program.md, measures the results of each iteration, and reverts non-improving edits using git to guide further improvements.