latex-arxiv-SKILL
A highly customizable agentic harness for arXiv-ready ML/AI review papers (and beyond). It drives agentic AI like Codex CLI and Claude Code through a gated LaTe
What is latex-arxiv-SKILL?
A highly customizable agentic harness for arXiv-ready ML/AI review papers (and beyond). It drives agentic AI like Codex CLI and Claude Code through a gated LaTeX workflow with verified BibTeX citations.
Overview of latex-arxiv-SKILL
What it is
The arXiv Review Paper Harness (also known as latex-arxiv-SKILL) is an agentic harness for writing machine-learning and AI review papers in LaTeX. It drives agentic AI tools such as Codex CLI and Claude Code through a gated LaTeX workflow with verified BibTeX citations. The harness turns a topic into an arXiv-ready ML/AI review paper that is gated, issue-driven, and verified citation by citation.
Key features
- Hard quality gates: no prose is written before the plan is approved, and no section is marked DONE until its acceptance criteria are met.
- Verified citations only: every
cite{}is web-checked against its source before being added toref.bib, preventing hallucinated references. - Issue-driven execution: an
issues/*.csvfile serves as the single source of truth for progress; the agent splits or inserts issues as scope grows. - Deterministic Python scripts: scaffolding, plan and issue generation, arXiv discovery (with a local SQLite cache), validation, and compilation are all Python scripts that behave the same on every run.
- Dual-runtime: the same skill bundle runs in both Claude Code and Codex.
- Compiles or it is not done: delivery requires a clean
pdflatexandbibtexbuild with no undefined-citation warnings.
Use cases
The harness is designed for writing review and survey articles for arXiv. It can also be used on an existing LaTeX project by running a citation-validation pass that audits and repairs ref.bib without re-scaffolding. With tailoring, the same gated workflow can extend to original or experimental research papers.
Getting started
See the repository README for installation and quick-start instructions.
License and pricing
The project is licensed under MIT and is open source (free).
Based on the project's public README and repository metadata.