autoresearch

Automate iterative problem solving by defining a Goal, Scope, and Metric to constrain the task, enabling the system to drive bounded loops, verify changes, and log results. It adapts to software, documentation, and planning workflows.

Autonomous task iteration loops through problem solving by constraining goals and measuring mechanical metrics. Verification works by running a defined Verify command and optional Guard after committing changes to ensure every iteration meets specified criteria.

Start an autonomous iteration loop by using the /autoresearch command with your plain-language goal and contextual data. The system generates a ready-to-run plan with inline configuration including Goal, Scope, Metric, and Verify parameters.

Yes, git-backed memory logs past experiments and results to enable learning from previous iterations. This memory supports subcommands like learn and predict to improve future autonomous planning and chained workflow execution.

Yes, autonomous iteration supports bounded and unbounded loops, interactive planning gates, and chainable workflows. Subcommands like plan, security, fix, scenario, and ship enable complex chained execution across software and documentation tasks.

Automated verification in unbounded loops requires a defined Verify command and optional Guard to prevent runaway iterations. Without mechanical metrics and bounded iterations, autonomous problem solving may lack convergence criteria for complex planning tasks.