Skip to content

Full Paper Writing Pipeline

Turns a research narrative into a reviewed, compiled LaTeX paper and PDF via chained sub-skills

Stars
18,482
Licence
MIT
Files
1 file

What it does

When installed, the agent runs a multi-stage workflow that takes a research narrative or notes and walks through planning an outline, generating figures, drafting LaTeX sections, compiling a PDF, and running review-and-revise rounds. It is built to sit on top of several other sub-skills named in the file (paper-plan, paper-figure, paper-write, paper-compile, and an improvement loop), and it can call external AI reviewer models through MCP tools. The visible file only defines the orchestration; the detailed behaviour lives in those other skills.

Great for

  • Researchers who already have experiment results and want a draft paper produced end to end
  • Students preparing a conference or journal submission who want structured help with LaTeX and revision
  • Teams that already use the sibling ARIS skill set and Codex or other reviewer MCP servers

Skip it if

  • Anyone expecting a finished, submission-ready paper without reviewing it: the output is AI-drafted and claims, citations and numbers must be checked by the author, since the file gives no verification step for facts.
  • Users without the sibling skills installed: the pipeline calls /paper-plan, /paper-write and others that are not in this file, so it will not run standalone.
  • Use with confidential or unpublished results on third-party reviewer models without checking data policies, because the review steps send paper content to external reviewer models.

Try asking

  • “Turn my NARRATIVE_REPORT.md into a full ICLR paper PDF”
  • “Write the paper pipeline from my experiment notes, targeting NeurIPS”
  • “Run the paper writing workflow with the style of this reference arXiv paper”
  • “Draft the paper from my research direction and results, then run two review rounds”
Step-by-step guide: how to use Full Paper Writing PipelineInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    paper-writing.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip paper-writing.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip paper-writing.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Provide a detailed NARRATIVE_REPORT.md with quantitative results and figure descriptions; the file says output quality depends on input detail.
  • Set AUTO_PROCEED to false or HUMAN_CHECKPOINT to true if you want to approve each phase or review round yourself.
  • Check every citation, number and claim in the final PDF before submitting, since the review loop is automated and AI-generated.

What's inside

  • SKILL.md

Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.