Run PaperOrchestra Research-Paper Skills
Use PaperOrchestra's host-agent skill pack to turn research ideas, logs, literature, figures, drafts, and autorater checks into a repeatable assisted LaTeX paper-writing pipeline.
npx skills add agentskillexchange/skills --skill run-paperorchestra-research-paper-skills
PaperOrchestra is useful when a researcher has scattered notes, experiments, figures, or prior agent work and wants a structured agent-assisted path toward a submission-ready LaTeX draft. The repeatable workflow is to use the PaperOrchestra skill pack with a capable host agent, prepare or aggregate idea and experimental-log inputs, run the outline, plotting, literature-review, section-writing, content-refinement, benchmark, and autorater skills, then inspect the generated artifacts and validation results.\n\nInvoke this skill when a user needs a controlled research-paper drafting workflow instead of asking a chat model to write a paper in one pass or editing manually from unstructured notes. Keep the scope to assisted drafting, transformation, and validation; do not present PaperOrchestra as a guarantee of research quality, a publication service, or the arXiv paper itself.
What this skill actually does
Inputs and prerequisites: PaperOrchestra skill pack, a host coding/research agent such as Claude Code, Cursor, Antigravity, Cline, Aider, OpenCode, or similar, local deterministic helper scripts, optional LaTeX and literature-search tools.
Setup notes: Clone the GitHub repository and point the host agent at the skills/paper-orchestra skill pack. Follow the repository’s host integration notes and run the deterministic helper scripts from the provided skills and references directories as instructed.
Source and verification boundary: use https://arxiv.org/pdf/2604.05018 as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.