This is the fourth step in the PaperOrchestra pipeline that generates a complete LaTeX manuscript in a single multimodal LLM call. It drafts the methodology, experiments, abstract, and conclusion sections, parses numeric results from experimental logs into booktabs tables, embeds figures generated in step 2, and merges everything into the template that already has your intro and related work. The implementation is careful about preserving existing content verbatim and includes deterministic gates to catch orphan citations and LaTeX syntax errors before passing to the refinement stage. It reads a TeX profile to know which packages are available so it generates compatible cross-references and formatting patterns. The single-call design is intentional for maintaining coherence across sections, and it uses vision input to ground figure references in what the images actually show.
npx -y skills add ar9av/paperorchestra --skill section-writing-agent --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
prisma/skills
firebase/agent-skills
wordpress/agent-skills
Dexploarer/hyper-forge
prisma/skills