peer-review

The peer-review skill serves as a rigorous, systematic framework for evaluating scientific manuscripts, research grant proposals, and technical documentation. Designed for researchers, journal editors, and academic reviewers, it transforms the subjective task of critiquing scientific work into an objective, checklist-driven process. By focusing on critical appraisal rather than simple editing, it ensures that every review addresses the core pillars of academic integrity: experimental rigor, statistical validity, reporting transparency, and clarity of communication. Users can leverage this skill to provide constructive, actionable feedback that helps authors strengthen their study design, address potential biases, and align with international standards.

• Performs deep-dive section analysis: assesses Abstract/Introduction for narrative impact and Methods/Results/Discussion for technical accuracy.

• Enforces strict compliance with established reporting guidelines, including CONSORT for clinical trials, STROBE for observational studies, and PRISMA for systematic reviews.

• Evaluates reproducibility by scrutinizing data availability, statistical method justification, sample size calculations, and control validity.

• Integrates with scientific-schematics to suggest and generate publication-quality diagrams, workflows, and decision trees to visualize research concepts.

• Identifies common research "red flags" such as over-interpretation of data, selective reporting, or circular analysis.

• This skill is highly effective when tasked with reviewing journal submissions, assessing grant feasibility, or refining manuscript drafts before final submission.

• For comprehensive evaluation of logical evidence or claim quality, pair this with the scientific-critical-thinking skill; for quantitative or rubric-based scoring, utilize the scholar-evaluation skill.

• Input expectations include the raw manuscript or proposal text; outputs provide a categorized, constructive critique structured by manuscript section.

• The agent is capable of providing feedback on statistical reporting, identifying missing error bars, ensuring proper use of p-values, and verifying the alignment between research hypotheses and experimental outcomes.