Academic paper review orchestration - Professor, Peer, and Official reviewers for scientific LaTeX papers
npx claudepluginhub mobled37/claude-labsAcademic paper review orchestration for Claude Code - CV Lab for computer vision papers
mobled37A Claude Code plugin for academic paper review orchestration. Features a multi-agent review pipeline and interactive Q&A with expert reviewers for computer vision LaTeX papers.
| Skill | Description |
|---|---|
CV Lab (cv-lab) | Full review pipeline - professor, 4 peers, revisions, official review loop |
Ask Professor (ask-professor) | Interactive Q&A with a senior CV professor about your paper |
Ask Peers (ask-peers) | Interactive Q&A with 4 specialized peer reviewers who search related papers |
# Add as a local marketplace
claude plugin marketplace add /path/to/claude_labs
# Install the plugin
claude plugin install claude-labs
CV Lab orchestrates a rigorous academic review loop with three types of reviewers:
| Reviewer | Model | Role |
|---|---|---|
| Professor | Opus | Defensive advisor review - tone, argument consistency, experimental rigor |
| 4 Peer Reviewers | Sonnet | Specialized parallel reviewers - architecture, math, figures, related works |
| Official Reviewer | Opus | Conference-style review with weighted 1-10 rating |
/claude-labs:cv-lab /path/to/paper_project/ # Default: threshold=7, max-rounds=5
/claude-labs:cv-lab . # Review project in current directory
/claude-labs:cv-lab /path/to/paper_project/ --threshold 8 # Require score 8+ (clear accept)
/claude-labs:cv-lab ./my-eccv-paper --threshold 6 --max-rounds 3 # Lenient, fewer rounds
/claude-labs:cv-lab /path/to/paper_project/ --threshold 5 --max-rounds 2 # Quick check for early drafts
| Parameter | Default | Description |
|---|---|---|
path | (required) | Path to the LaTeX project directory (containing main.tex, sections, figures, bib, etc.) |
--threshold N | 7 | Minimum official review score (1-10) to accept |
--max-rounds M | 5 | Maximum review-revise cycles |
The skill auto-discovers the project layout. A typical CV paper project:
paper_project/
main.tex # Root .tex file (entry point)
main.bib # Bibliography
src/
sec/
1. introduction.tex # Section files
2. related works.tex
3. preliminaries.tex
4. proposed method.tex
5. experiments.tex
6. conclusion.tex
x_supplementary.tex
imgs/
figure1.pdf # Figures
...
eccv.sty # Style files (auto-detects venue)
llncs.cls # Document class
Any layout is supported: flat, nested, or multi-file with \input{}/\include{}.
+-------------------+
| Professor Review |
| (tone, arguments, |
| experiments) |
+---------+---------+
|
v
+---+-----------+-----------+---+
| | | | |
v v v v |
Peer-1 Peer-2 Peer-3 Peer-4 |
Arch. Math Figures Related |
Works |
+---+-----------+-----------+---+
|
v
+---------+---------+
| Plan & Revise |
| (aggregate and |
| fix issues) |
+---------+---------+
|
v
+---------+---------+
| Official Review |
| (1-10 rating) |
+---------+---------+
|
score >= threshold?
/ \
Yes No
| |
v v
ACCEPT Loop back to
Professor Review
The professor acts as a protective advisor who wants the paper to succeed at a top venue. They challenge every weakness before external reviewers find them: