Paper Orchestra

Implements the Song et al. 2026 PaperOrchestra pipeline: five coordinated agents that transform raw experimental logs and a research idea into a conference-ready LaTeX paper with compiled PDF. The orchestrator runs outline generation, then parallelizes figure plotting and literature review (with Semantic Scholar integration), reconciles citations back into the outline, writes sections, and iteratively refines content. Ships with pre-flight validation scripts that check input density, TeX package availability, and idea-to-experiment consistency before burning tokens. The anti-leakage prompt from Appendix D.4 is mandatory across all LLM calls to keep outputs grounded in your data rather than regurgitating training set papers. Expect 50-70 LLM calls for a full run.

GitHub stars: 572
last updated: Jun 2, 2026

View on GitHub

Install

$ npx -y skills add ar9av/paperorchestra --skill paper-orchestra --agent claude-code

Installs into .claude/skills of the current project.

Files

paper-orchestra (Orchestrator)

Top-level driver for the PaperOrchestra pipeline. Read this document and follow the steps below. The detailed prompts and rules live in each sub-skill's SKILL.md and references/ directories — you (the host agent) will load them as you go.

Source paper: Song et al., PaperOrchestra: A Multi-Agent Framework for Automated AI Research Paper Writing, arXiv:2604.05018, 2026. https://arxiv.org/pdf/2604.05018

What this skill produces

A complete submission package P = (paper.tex, paper.pdf) written into workspace/final/, plus a full audit trail under workspace/ (outline, figures, refs, drafts, refinement worklog, provenance snapshot).