---
name: reviewerzero
description: Automated LaTeX citation and numeric claim auditing system for IEEE and academic research papers. Use when analyzing .tex and .bib source files, verifying numerical claims against cited references, resolving papers via arXiv/Crossref/Semantic Scholar, or preventing citation drift before paper submission.
---

# ReviewerZero: Academic Paper & Citation Claim Auditor

**ReviewerZero** parses `.tex` source and `.bib` bibliography files, extracts numerical claims, percentages, and units near citation keys, resolves referenced literature via arXiv, Crossref, and Semantic Scholar, and cross-checks claims against resolved full-text sources.

## Core Capabilities

1. **LaTeX & BibTeX Parsing**: Extracts citation keys (`\cite{...}`) and surrounding numeric context.
2. **Online & Offline Source Resolution**: Resolves papers via arXiv ID, DOI, or Semantic Scholar API. Supports local PDF directories (`--sources`).
3. **Numeric Claim Matching**: Evaluates claim tolerance (default `tight=0.5%`, `loose=5.0%`) and flags mismatches or missing citations.
4. **HTML/CLI Reporting**: Generates interactive HTML audit reports (`audit.html`) highlighting verified, close match, and mismatched claims.

## Trigger Scenarios

Activate this skill when:
- Reviewing or auditing a LaTeX paper (`.tex` / `.bib` files).
- Checking paper citations for accuracy or numerical consistency.
- Verifying reported experimental metrics against cited baseline papers.
- Running paper claim verification before academic venue submission.

## Usage Commands

```bash
# Run paper audit on paper.tex with refs.bib
python -m reviewerzero.cli paper.tex --bib refs.bib --out audit.html

# Run offline audit using local source PDFs
python -m reviewerzero.cli paper.tex --bib refs.bib --sources ./sources/ --out audit.html
```

## Guidelines

- Ensure `.tex` and `.bib` files exist in the specified workspace path.
- Review flagged mismatches in `audit.html` for potential rounding errors or unit discrepancies.