AgentSkillsCN

paper-reviewer

全面的投稿前审核统筹者。协调方法学检查、偏见检测、AI 文本识别、统计验证，以及期刊投稿的报告指南合规性检查。

SKILL.md
--- frontmatter
name: paper-reviewer
description: Comprehensive pre-submission review orchestrator. Coordinates methodology checks, bias detection, AI text detection, statistical validation, and reporting guideline compliance for journal submissions.
version: 1.0.0
author: Aaron Storey

Paper Reviewer

Comprehensive pre-submission review for journal manuscripts. Orchestrates multiple specialized skills from research-assistant to ensure publication readiness.

When to Use

  • •Before submitting papers to journals
  • •After completing systematic reviews
  • •When preparing dissertation chapters for committee
  • •For quality assurance on research manuscripts

Features

  1. •AI Text Detection - Identify LLM-generated patterns
  2. •Methodology Assessment - Verify research design rigor
  3. •Bias Detection - Apply RoB 2, ROBINS-I, QUADAS-2 frameworks
  4. •Statistical Validation - Check assumptions, effect sizes, power analysis
  5. •Reporting Compliance - Validate CONSORT, PRISMA, STROBE guidelines
  6. •Citation Integrity - Verify references, check for retractions
  7. •Reproducibility Audit - Assess code/data availability

Usage

code
Review the manuscript at ~/Documents/submissions/AffectivePrompting/ for journal submission.
This is a systematic review targeting IEEE Transactions.

Review Pipeline

Phase 1: Triage

  • •Classify paper type (RCT, systematic review, observational, etc.)
  • •Identify target journal requirements
  • •List available files

Phase 2: AI Detection

  • •Run ai-check skill
  • •Flag sections with AI patterns >30%
  • •Generate authenticity report

Phase 3: Methodology

  • •Assess study design appropriateness
  • •Verify protocol adherence
  • •Check randomization/blinding (if applicable)

Phase 4: Risk of Bias

  • •Apply appropriate RoB tool
  • •Document domain-by-domain assessment
  • •Generate traffic light visualization

Phase 5: Statistical Validation

  • •Verify assumption testing
  • •Check effect size reporting
  • •Validate power analysis

Phase 6: Reporting Compliance

  • •Apply appropriate checklist (CONSORT/PRISMA/STROBE)
  • •Calculate compliance percentage
  • •List missing items

Phase 7: Citation Audit

  • •Verify all references
  • •Check for retractions
  • •Validate DOIs

Phase 8: Reproducibility

  • •Check code availability
  • •Verify data accessibility
  • •Assess documentation completeness

Output

Generates comprehensive review report:

  • •Executive summary with overall assessment
  • •Section-by-section findings
  • •Priority-ranked revision checklist
  • •Detailed appendices

Paper Type-Specific Workflows

Systematic Reviews

  • •PRISMA 2020 compliance
  • •AMSTAR 2 quality assessment
  • •Meta-analysis validation

RCTs

  • •CONSORT compliance
  • •RoB 2 bias assessment
  • •ITT/per-protocol analysis check

Observational Studies

  • •STROBE compliance
  • •ROBINS-I bias assessment
  • •Confounding control verification

Integration

Uses research-assistant skills:

  • •ai-check
  • •risk-of-bias
  • •hypothesis-test
  • •effect-size
  • •power-analysis
  • •publication-prep
  • •citation-format
  • •prisma-diagram

Configuration

Default thresholds (can override):

  • •AI detection: 30% (flag if higher)
  • •Reporting compliance: 85% minimum
  • •All 8 phases enabled by default

Part of ai_scientist research workflow automation