Paper Revision and Polishing Full-Dimension Dashboard

Addressing Systematic Challenges in Paper Revision

After receiving peer review comments, researchers often struggle to systematically process feedback, diagnose issues, and execute effective revisions. Manual approaches can lead to missed key points or overlooked structural imbalances, especially when dealing with traces of AI-generated content, potentially reducing paper readability or journal compliance. The Paper Revision and Polishing Full-Dimension Dashboard provides a five-phase pipeline that automates the entire workflow from comment analysis to final delivery, minimizing human error and enhancing revision quality.

Core Workflow and Key Capabilities

Based on the SOP in SKILL.md, the skill operates through five phases:

• Phase 0: Manuscript Comment Parsing: Automatically extracts actionable items from review comments, categorizes them by risk level (🔴 Critical, 🟠 Suggested, 🔵 Polishing), and outputs an action checklist. It supports input from text, screenshots, or PDFs, while extracting journal format requirements.

• Phase 1: Full-Spectrum Diagnosis: Runs the diagnose_paper.py script to obtain quantitative metrics like word count, sentence length, and AI Flavor Density Index. It loads the ai-flavor-patterns.md lexicon to detect five issue types (critical to highlights), generating a diagnostic report with structural imbalances and readability problem areas.

• Phase 2: Visual Annotation Generation: Based on diagnosis results and review comments, it uses an HTML template to create an interactive annotation report. Each paragraph includes original highlighting, issue diagnosis, and rewriting examples, color-coded by problem type, aiding authors in understanding revision logic.

• Phase 3: Graded Rewriting Execution: Performs modifications in A → B → C priority order. Class A focuses on clearing AI traces, e.g., replacing specific tool names with functional categories; Class B optimizes sentence structures and flow; Class C involves polishing. Principles emphasize preserving empirical data, adding concrete details, and controlling paragraph density.

• Phase 4: Final Review Self-Check: Uses the quality-checklist.md checklist to confirm items, generating a self-check report. It determines if the paper is deliverable, needs further revision, or requires rewriting, ensuring compliance with standards.

Additionally, the Enhanced Report Generator offers a 15-dimension, three-level interactive dashboard, covering aspects from overview to final checklist, with features like paragraph expansion, tool name replacement suggestions, and revision session tracking, suitable for in-depth analysis.

Applicable Boundaries and Considerations

• This skill relies on sub-skills like humanizer, research-writing, and academic-figures, which are invoked automatically during the process, but these must be installed.
• Input requires original review comments (text, screenshots, or PDF); otherwise, Phase 0 may not execute properly.
• It is designed for academic paper revision, particularly for journal submissions, but may need manual adjustments for specific fields or formats.
• During rewriting, the humanizer skill focuses on removing AI traces, but for highly specialized content, domain knowledge verification is recommended.