跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
设计与 UI

paper-writing

Guides writing academic papers section by section using an 11-step workflow with LaTeX templates and counterintuitive writing tactics. Covers Abstract, Introduction, Method, Experiments, Related Work, Conclusion, and Supplementary. Use when: user asks to write or draft a paper section, needs LaTeX templates, wants to improve academic writing quality, optimize novelty framing, or mentions 'write introduction', 'draft method', 'paper writing'. Do NOT use for pre-submission review (use paper-review), experiment execution (use experiment-pipeline), or paper planning/story design (use paper-planning).

CamusGIT · 已有 AI 审核记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add CamusGIT/EvoQuant --skill paper-writing

技能说明


name: paper-writing description: "Guides writing academic papers section by section using an 11-step workflow with LaTeX templates and counterintuitive writing tactics. Covers Abstract, Introduction, Method, Experiments, Related Work, Conclusion, and Supplementary. Use when: user asks to write or draft a paper section, needs LaTeX templates, wants to improve academic writing quality, optimize novelty framing, or mentions 'write introduction', 'draft method', 'paper writing'. Do NOT use for pre-submission review (use paper-review), experiment execution (use experiment-pipeline), or paper planning/story design (use paper-planning)." allowed-tools: "write_file edit_file read_file think_tool" metadata: author: EvoQuant version: '1.0.0' tags: [core, research, writing, academic-writing, latex]

Paper Writing

A systematic 11-step workflow for writing academic papers, with section-specific templates and battle-tested writing principles.

When to Use This Skill

  • User asks to write or draft a paper or paper section
  • User needs LaTeX templates for Abstract, Introduction, Method, Experiments, etc.
  • User wants to improve academic writing quality
  • User mentions "paper writing", "write introduction", "draft method section", etc.

Artifact Sources

If you used upstream skills earlier in the pipeline, pull these artifacts before writing:

Source SkillArtifactUsed In
paper-planningStory summary (task → challenge → insight → contribution → advantage)Steps 1-2 (Introduction writing plan)
paper-planningModule Motivation Mapping tableStep 3 (Method subsections)
paper-planningExperiment plan (comparisons + ablations + demos)Step 5 (Experiments section)
paper-planningPipeline figure sketchSteps 1, 6 (Method overview figure)
paper-planningClaim-to-experiment mappingSteps 2, 5 (Abstract, Introduction, Experiments)
paper-planningFallback narrative (if planned)Steps 7-8 (Introduction / Conclusion pivot)
experiment-pipelineStage 1-4 results, ablation tables, trajectory logsStep 5 (write experiments)
experiment-craftFailure analysis, implementation tricksStep 3 (Method section), Step 9 (limitations)

The 11-Step Writing Process

Follow these steps in order. Each step builds on the previous one.

  1. Draw a pipeline figure sketch — Sketch the method's pipeline figure to clarify the overall approach. The figure highlights novelty, not just explanation.
  2. Design the story and plan experiments — Outline the paper's story (core contribution, module motivations). List comparison experiments and ablation studies. Draft an Introduction writing plan.
  3. Write Method — Organize the Method writing plan, then draft Method. Run experiments in parallel.
  4. Revise Introduction and Method — Iterate on both sections while experiments continue.
  5. Write Experiments — Once experiments are mostly done, organize the Experiments writing plan, then draft.
  6. Polish figures — Finalize the pipeline figure. Create the teaser figure.
  7. Write Related Work — List related papers, group into topics, write paragraphs.
  8. Review the paper — Self-review Introduction, Method, and Experiments. Use the paper-review skill.
  9. Write Abstract — Organize the Abstract writing plan, then draft.
  10. Choose the title — List important keywords, then compose an informative title.
  11. Iterate — Repeatedly review and revise the entire paper.

Counterintuitive Writing Rules

Apply these rules when aiming for higher acceptance probability:

  1. Underclaim in prose, overdeliver in evidence: Reduce adjective intensity in Abstract/Introduction; let tables and figures carry the strength.
  2. State one meaningful limitation early: A controlled limitation statement increases credibility and lowers reviewer suspicion.
  3. Lead with mechanism, not only metric: Explain why the method works before listing numbers; reviewers trust causal logic more than isolated gains.
  4. Prefer one decisive figure over many average figures: Build one "cannot-ignore" figure that validates the central claim under hard conditions.
  5. Remove weak but flashy claims: Any claim without direct evidence should be deleted, even if it sounds impressive.
  6. Declare scope boundaries explicitly: One sentence in Introduction and Conclusion stating what your method targets reduces reviewer fear of hidden assumptions.
  7. Show one failure case: Include one representative failure with diagnosis — it signals competence, not weakness.

See references/counterintuitive-writing.md for all 7 tactics with before/after examples.

Section Quick Reference

Abstract

Answer these questions before drafting:

  1. What technical problem do we solve, and why is there no well-established solution?
  2. What is our technical contribution?
  3. Why does our method fundamentally work?
  4. What is our technical advantage / new insight?

Three template versions: challenge-first, insight-bridge, multi-contribution. See references/abstract-templates.md

Introduction

Thinking process (reverse then forward):

  • Reverse: (1) What is the technical problem? (2) What are our contributions? (3) Benefits and new insights? (4) How to lead into the challenge?
  • Forward: (1) Task → (2) Previous methods → challenge → (3) Our contributions → (4) Technical advantages and insights

Four ways to introduce the task, three ways to present challenges, four ways to describe the pipeline. See references/introduction-templates.md

Anti-pattern: Never write "here is a naive solution, then our improvement" — this makes the work appear incremental.

Method

Every pipeline module needs three elements:

  1. Module design — Data structure, network design, forward process (given X inpu
阅读完整来源 ↗