auto-review-loop

Run multi-round artifact reviews with fixes until positive assessment or max rounds.

Updated Jun 10, 2026
One-click install
npx skills add https://github.com/xqinag/ARIS-new --skill auto-review-loop-xqinag
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: auto-review-loop
Source: https://github.com/xqinag/ARIS-new/tree/main/skills/auto-review-loop
Command: npx skills add https://github.com/xqinag/ARIS-new --skill auto-review-loop-xqinag

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Autonomous multi-round research review loop that repeatedly reviews artifacts, implements fixes, and re-reviews until a positive assessment or max rounds are reached.

Core Features & Use Cases

  • Automated, iterative peer review: runs rounds, tracks memory, and stops on positive verdict
  • Configurable backends: supports codex/manual reviewer backends with memory and debate protocols
  • Traceable outputs: logs to review-stage/AUTO_REVIEW.md and REVIEW_STATE.json for audit and resume

Quick Start

Start an autonomous review loop on a given topic by invoking the skill with the topic argument.

Frequently Asked Questions about auto-review-loop

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

FAQPage Schema
How do I automate iterative peer review for research papers?

Automated iterative peer review runs multiple rounds of artifact evaluation, proposes fixes, and re-reviews until a positive assessment or max rounds are reached. This loop applies to research papers, code reviews, and experimental results.

What is an autonomous review loop and when do I need it?

An autonomous review loop is a process that repeatedly evaluates research artifacts, implements fixes, and re-reviews them. You need it when a project requires iterative improvement cycles with full traceability for auditability.

Can I use a manual reviewer backend instead of codex for automated research reviews?

Yes, the autonomous review loop supports configurable backends including both codex and manual reviewer modes. Both options feature memory tracking and debate protocols to evaluate artifacts across multiple rounds.

How are review states logged and resumed in an automated research review workflow?

Review states are logged to REVIEW_STATE.json and detailed outputs are written to review-stage/AUTO_REVIEW.md. These files provide full traceability, enabling you to audit the review history and resume interrupted review loops.

What are the limitations of autonomous multi-round research review loops?

Autonomous multi-round research review loops are bounded by a maximum round limit to prevent infinite cycles. If the reviewer cannot reach a positive assessment within the configured max rounds, the loop terminates with the final unresolved state logged.

Does the automated review loop work for code reviews and experimental results?

Yes, the automated review loop evaluates code reviews and experimental results alongside research papers. It identifies weaknesses, proposes fixes, and re-evaluates any project artifact requiring iterative improvement and traceability.