Agent Skills

Peer Review

AIPOCH

Conduct professional peer reviews for papers or theses, providing structured evaluations and improvement suggestions; use when you need a pre-submission assessment, an internal review, or academic quality control.

4
0
FILES
peer-review/
skill.md
references
guide.md
assets
peer_review_template.md
86100Total Score
View Evaluation Report
Core Capability
84 / 100
Functional Suitability
11 / 12
Reliability
9 / 12
Performance & Context
7 / 8
Agent Usability
14 / 16
Human Usability
8 / 8
Security
10 / 12
Maintainability
9 / 12
Agent-Specific
16 / 20
Medical Task
20 / 20 Passed
92Pre-submission manuscript check: Before submitting to a journal/conference to identify major risks and revision priorities
4/4
88Internal lab/group review: For advisor or team quality control prior to external dissemination
4/4
86Structured end-to-end review workflow: Overall evaluation → methods/results check → issue organization → recommendation
4/4
86Major vs. minor issue triage: Separates publication-blocking problems from polish-level improvements
4/4
86End-to-end case for Structured end-to-end review workflow: Overall evaluation → methods/results check → issue organization → recommendation
4/4

SKILL.md

Peer Review

When to Use

  • Pre-submission manuscript check: Before submitting to a journal/conference to identify major risks and revision priorities.
  • Internal lab/group review: For advisor or team quality control prior to external dissemination.
  • Thesis/dissertation evaluation: To assess academic rigor, structure, and defensibility before committee review.
  • Revision planning after feedback: To translate reviewer/editor comments into an actionable improvement roadmap.
  • Quality assurance for research outputs: To ensure methods, reporting, and conclusions meet disciplinary standards.

Key Features

  • Structured end-to-end review workflow: Overall evaluation → methods/results check → issue organization → recommendation.
  • Major vs. minor issue triage: Separates publication-blocking problems from polish-level improvements.
  • Actionable revision suggestions: Each issue is paired with concrete steps to fix or strengthen the work.
  • Recommendation with rationale: Clear accept/revise/reject guidance with reasons and improvement path.
  • Reusable templates and checklists: Supports consistent formatting and comprehensive coverage (see referenced files).

Dependencies

  • None (runtime)

Example Usage

Use the template to produce a structured review.

  1. Open the template:

    • assets/peer_review_template.md
  2. Fill it using the workflow below. Example (copy/paste and complete):

# Peer Review Report

## 1. Overall Evaluation
**Summary of the work:**  
This paper investigates [research question] by using [method/data]. The main contributions are: (1) [...], (2) [...].

**Novelty and significance:**  
- Novelty: [high/medium/low] because [...]
- Significance: [high/medium/low] because [...]

## 2. Methods and Results
**Research design and methodology:**  
- Appropriateness of design: [...]
- Data and sampling: [...]
- Statistical/analytical methods: [...]
- Reproducibility (code/data availability, parameter reporting): [...]

**Results vs. conclusions:**  
- Do results support claims? [...]
- Alternative explanations addressed? [...]
- Robustness checks/ablation/sensitivity analysis: [...]

## 3. Issues and Revision Suggestions

### Major Issues (must address)
1. **Issue:** [...]
   - **Why it matters:** [...]
   - **Suggested fix:** [...]
   - **Expected impact:** [...]

2. **Issue:** [...]
   - **Why it matters:** [...]
   - **Suggested fix:** [...]
   - **Expected impact:** [...]

### Minor Issues (should address)
1. **Issue:** [...]
   - **Suggested fix:** [...]

2. **Issue:** [...]
   - **Suggested fix:** [...]

## 4. Recommendation
**Recommendation:** Accept / Minor Revision / Major Revision / Reject

**Rationale:**  
Explain the decision based on novelty, rigor, clarity, and evidence strength.

**Path to improvement:**  
List the top 3–5 changes that would most improve the manuscript.

For output formats, checklists, and inspection points, see:

  • references/guide.md

Implementation Details

Review Workflow (Algorithm)

  1. Read for global understanding

    • Read the abstract and full text to form an overall impression.
    • Identify the research question, claimed contributions, and target audience/venue.
  2. Overall evaluation

    • Summarize the research questions and major contributions.
    • Assess novelty (what is new vs. prior work) and significance (why it matters).
  3. Methods and results verification

    • Check research design, data quality, and statistical/analytical methods for correctness and suitability.
    • Evaluate whether results logically and quantitatively support the conclusions.
    • Flag missing details that prevent replication (e.g., parameters, datasets, baselines, evaluation protocol).
  4. Issue organization

    • Classify findings into:
      • Major issues: validity threats, methodological flaws, unsupported claims, missing critical experiments, ethical/compliance gaps.
      • Minor issues: clarity, formatting, citations, small inconsistencies, language improvements.
    • For each issue, provide an actionable revision suggestion (what to change and how).
  5. Recommendation

    • Provide a decision (accept/revise/reject) aligned with the severity and fixability of major issues.
    • Explain the rationale and provide a prioritized improvement path.

Key Parameters / Criteria

  • Novelty: degree of differentiation from prior work; clarity of contribution statement.
  • Significance: practical/theoretical impact; relevance to the field and venue.
  • Rigor: appropriateness of methods; correctness of analysis; robustness checks.
  • Evidence alignment: strength of support from results to claims; avoidance of overgeneralization.
  • Reproducibility: completeness of experimental details; availability of data/code; transparent reporting.
  • Clarity and structure: logical flow, readability, figure/table quality, and citation completeness.

Templates and References

  • Template (preferred for structured output): assets/peer_review_template.md
  • Guidance/checklists/output formats: references/guide.md