open-reviewer

Validate abstract completeness, length, and alignment.

Enforce evidence-grounded outputs with explicit uncertainty.

Screen for excessive self-citation and suspicious citation patterns.

Run contradiction and alignment checks across all sections.

Check interpretation quality, limitations, and overclaiming.

Create stable anchors for sections, paragraphs, figures, tables, equations, and references.

Review datasets, baselines, metrics, fairness, and leakage risks.

Review figures and tables for caption quality, linkage, and alignment.

Normalize issues into a single production schema.

Checklist for grammar, clarity, terminology, and abbreviations.

Check equations, notation, and symbol definitions.

Generate final scorecard, recommendation, and executive summary.

Review assumptions, completeness, logic, and design rationale.

Evaluate contribution clarity, distinctness, and non-triviality.

Normalize raw paper inputs into a stable parsed structure.

Aggregate ethical, integrity, and suspicious-pattern warnings.

Validate reference list quality and citation/reference alignment.

Review related work accuracy, relevance, and gap legitimacy.

Verify that results support claims and surface red flags.

Format section-wise findings into a consistent review payload.

Assign issue severity consistently.

Validate section order, transitions, hierarchy, and caption linkage.