autoresearch

Iterate autonomously on a defined objective through modify-verify-keep/discard loops.

Updated Aug 23, 2026

Search Tags:#automation #metrics #iteration #autonomy #loop #autoresearch

Ready to run

npx skills add https://github.com/johngrandson/orkestry --skill autoresearch-johngrandson

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autoresearch
Source: https://github.com/johngrandson/orkestry/tree/main/.claude/skills/autoresearch
Command: npx skills add https://github.com/johngrandson/orkestry --skill autoresearch-johngrandson

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Autonomous goal-directed iteration to continuously improve tasks by looping through modify-verify-keep/discard until termination.

Core Features & Use Cases

Autonomous looping and decision-making based on measurable criteria
Constraint-driven iteration with safety and rollback guidance
Applies to code, content, design and data tasks with clear metrics

Quick Start

Invoke the autoresearch loop on a defined task to start continuous iteration with automatic evaluation.

Frequently Asked Questions about autoresearch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate iterative code generation with continuous refinement?▼

Automate iterative code generation by running an autonomous loop that continuously refines a defined objective through modify, verify, and keep or discard cycles. The loop evaluates mechanical verification metrics to guide each subsequent iteration automatically.

What is autonomous goal-directed iteration for content optimization?▼

Autonomous goal-directed iteration is a looping process that continuously improves content optimization tasks by applying one change at a time. It uses measurable criteria to evaluate modifications, keeping improvements and discarding failures until termination conditions are met.

Can I use autonomous loops for design exploration with bounded constraints?▼

You can use autonomous loops for design exploration within bounded constraints. The loop applies safety guidance and rollback mechanisms to ensure each experimental change is mechanically verified before being kept or discarded in the persistent results log.

How do I start an autonomous looping task for persistent improvement?▼

Start an autonomous looping task by invoking the loop on a defined objective. The system automatically handles decision-making based on measurable criteria, executing one change at a time and logging persistent results to guide next iterations without manual intervention.

Does autonomous iteration work without external dependencies?▼

Autonomous iteration works without external dependencies. The loop operates independently by relying on its internal mechanical verification logic, constraint-driven safety guidance, and a persistent results log to drive the modify, verify, and keep or discard workflow.

What are the limitations of constraint-driven iteration for autonomous tasks?▼

Constraint-driven iteration is limited by the need for clear metrics and mechanical verification. Without measurable criteria and bounded constraints, the autonomous loop cannot effectively evaluate modifications, make decisions, or guide the next iterations for code, content, or data tasks.