auto-review-loop

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.

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.

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.

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.

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.

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.