academic-paper-reviewer

Simulates a five-reviewer peer review panel producing editorial decisions and revision roadmaps.

Updated Apr 27, 2026
One-click install
npx skills add https://github.com/DUT-AI/intelligent-testing --skill academic-paper-reviewer-dut-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: academic-paper-reviewer
Source: https://github.com/DUT-AI/intelligent-testing/tree/main/.agents/skills/academic-paper-reviewer
Command: npx skills add https://github.com/DUT-AI/intelligent-testing --skill academic-paper-reviewer-dut-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Authors lack access to realistic pre-submission peer review, so manuscripts get desk-rejected or hit with avoidable major revisions. This Skill simulates a full journal review process with five independent reviewer personas, surfacing methodological, domain, cross-disciplinary, and adversarial weaknesses before submission. ## Core Features & Use Cases - Multi-Perspective Panel Review: A field analyst configures five reviewers (Editor-in-Chief, methodology, domain, perspective, and Devil's Advocate) who independently review the paper, followed by an editorial synthesis producing a decision letter and prioritized revision roadmap. - Six Operational Modes: full review, re-review (revision verification with a traceability matrix), quick assessment, methodology-focus, Socratic guided review, and calibration mode that measures the reviewer's own FNR/FPR against a user-supplied gold set. - Use Case: Before submitting a manuscript on higher education policy, run a full review to receive five independent reports, an editorial decision, and a revision roadmap that feeds directly into a paper-revision workflow. ## Quick Start Ask the assistant to review your paper by pasting the manuscript text or providing the file, optionally specifying a mode such as quick, methodology-focus, or re-review.

Frequently Asked Questions about academic-paper-reviewer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a simulated peer review of my paper before submission?

Paste your manuscript or provide the file and request a review. The skill identifies the paper's field, configures five reviewer personas, runs independent reviews, and returns an editorial decision letter with a prioritized revision roadmap.

What review modes does the academic paper reviewer support?

It supports six modes: full panel review, re-review for verifying revisions, quick 15-minute assessment, methodology-focus, Socratic guided review, and calibration mode that measures reviewer accuracy against a labeled gold set of papers.

Can the reviewer verify that my revisions addressed earlier comments?

Yes, re-review mode takes the original revision roadmap, the revised manuscript, and an optional response letter, then produces a verification report with a traceability matrix, residual issues, and a new editorial decision.

Does the AI reviewer modify my manuscript during review?

No. The skill enforces a strict read-only constraint: reviewers produce separate reports, decisions, and roadmaps but never edit the submitted manuscript. Revision suggestions are delivered as documents for the author to apply.

How accurate are the AI review scores compared to human reviewers?

Accuracy is measurable through calibration mode, which runs the full review five times per paper on a user-supplied gold set of 5-20 labeled papers and reports FNR, FPR, balanced accuracy, and per-dimension calibration error with confidence intervals.

What are the limitations of simulated peer review?

Calibration profiles are domain-specific and session-scoped, so results do not transfer across fields or sessions. The reviewer can also misjudge severity on field-norm issues, which is why critical findings require externally grounded evidence.