What problem does it solve?
It helps you iteratively refine a research draft by running a structured review-improve-review loop with an external LLM, reducing the risk of submitting work that still has major weaknesses.
Core Features & Use Cases
- Autonomous review loop: Performs repeated cycles of reviewer assessment, fix implementation, and re-review until quality passes or the maximum rounds are reached.
- Action-driven improvements: Extracts score, verdict, and prioritized weaknesses from reviewer output, then applies the minimum necessary fixes.
- Persistent recovery and logging: Saves compact round state to REVIEW_STATE.json and appends a cumulative, expandable review history to AUTO_REVIEW.md for continuity across runs.
- Use Case: You have a draft with experiments and results; this Skill reviews it like a senior venue reviewer, identifies critical gaps, applies targeted changes, and rechecks readiness across up to four rounds.
Quick Start
Run auto review loop llm on your research topic to produce a score, verdict, and prioritized fix plan, then continue improving until the work is ready or MAX_ROUNDS is reached.
