autoresearch

Run a stateful research loop with artifact-gated validation.

6|Updated Apr 26, 2026

Search Tags:#research automation #research loop #stateful validation #continuous research #validation artifact

Ready to run

npx skills add https://github.com/Weihong-Liu/GenericAgent_project --skill autoresearch-weihong-liu

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autoresearch
Source: https://github.com/Weihong-Liu/GenericAgent_project/tree/main/.codex/skills/autoresearch
Command: npx skills add https://github.com/Weihong-Liu/GenericAgent_project --skill autoresearch-weihong-liu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Automates the research process with persistent validation and artifact-gated completion, reducing manual oversight and ensuring quality.

Core Features & Use Cases

Stateful Workflow: Continuously nudges research forward until explicit validation evidence is present.
Validation Mode Choice: Supports both script-based and prompt-based validation for flexibility.
Artifact-Gated Completion: Ensures research loop completion is based on validated outcomes.
Use Case: Ideal for long-term research projects requiring continuous refinement and validation.

Quick Start

To initiate a research loop with script validation, use the command: deep-interview --autoresearch to define the mission and validator, then run autoresearch.

Frequently Asked Questions about autoresearch

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

FAQPage Schema

How do I automate continuous research with stateful validation?▼

Automate continuous research by configuring a stateful loop that continuously nudges the process forward until explicit validation evidence passes artifact-gated completion criteria. This reduces manual oversight by persistently checking research outcomes against defined validators.

What is artifact-gated completion in a research loop?▼

Artifact-gated completion ensures a research loop finishes only when validated outcomes are present. It requires explicit validation evidence from either an external script or prompt-based validator before allowing the continuous research process to successfully complete.

How do I set up script-based validation for a research automation loop?▼

Set up script-based validation by defining your mission and validator script using the `deep-interview --autoresearch` command, then run `autoresearch`. The stateful loop uses this external validator script to validate artifacts and trigger completion.

Does research automation support both script and prompt-based validation?▼

Yes, research automation supports both script-based and prompt-based validation methods. This validation mode choice provides flexibility, allowing you to choose between executing an external validator script or using prompt-based criteria to gate artifact completion.

When should I use a persistent research loop for my projects?▼

Use a persistent research loop for long-term research projects requiring continuous refinement and validation. It is ideal when you need automated nudging and stateful validation to ensure quality outcomes without manual intervention.

What do I need to run a stateful research automation loop?▼

You need an external validator script or prompt and a persistent state management system to run a stateful research automation loop. These dependencies are required to enforce artifact-gated completion and continuous validation.