What problem does it solve?
Preparing research for top-tier venue submission requires repeated rounds of review, issue identification, fix implementation, and re-review, which is labor-intensive and prone to oversight when managed manually.
Core Features & Use Cases
- Autonomous Multi-Round Review Loop: Automatically coordinates external reviewer assessments (via Codex or manual backends), implements fixes for identified weaknesses, and re-reviews until a positive assessment is received or the maximum round limit is hit.
- Configurable Review Settings: Supports adjustable reviewer difficulty levels, optional human checkpoints for fix approval, and state persistence to resume long-running loops after context compaction.
- Structured Review Documentation: Automatically logs all review rounds, raw reviewer feedback, debate transcripts, and implemented fixes to a cumulative review file for full auditability.
Use Case: A researcher preparing a MICCAI 2025 paper can use this skill to automatically iterate through review cycles, address weaknesses like missing ablations or unclear method descriptions, and track progress toward submission readiness without manual coordination.
Quick Start
Invoke the auto-review-loop skill with your research project's scope to start an autonomous iterative review and fix cycle that runs until the work receives a positive assessment from the external reviewer or reaches the maximum round limit.
