SkillAtlasSkill 详情

academic-paper-composer

Security audit: baseline 52/52 CLEAN

审核状态:已审核Quality 80Security 80

复制安装命令

用 Codex 或 Claude 安装复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它先审查 Skill 页面再帮你安装。

复制前请先查看来源、License 和安全提示。

项目 README

来源文件:README.md

抓取于 2026年8月4日

Awesome GitHub stars License: CC BY-SA 4.0 PRs Welcome Validate catalog OpenSSF Scorecard Security audit: baseline 52/52 CLEAN Rigor coverage Powered by StatsPAI

Auto-Empirical Research Skills (AERS)

📌 文档结构(2026-07-22 起): 本文件是中文默认入口 —— banner + badges + 信任面 + 9 阶段流水线速览 + 76 行合集总表。 每个合集的完整描述、按用途分组、精确数字、验证方法在 docs/CONTENT_ZH.md(扩展正文,总表行内的 → 直接跳转到对应锚点)。

English version: README-en.md · 中文扩展正文:docs/CONTENT_ZH.md · README-zh-CN.md 已弃用(重定向占位)

🌐 语言: English | 简体中文(默认) | 繁體中文 | 日本語 | 한국어


CoPaper.AI Stanford REAP - Center on China's Economy & Institutions

Stanford REAP × CoPaper.AI · 实证研究 AI 工具的学术工业级产品
由斯坦福实证研究方法论团队打造,覆盖从数据清洗到顶刊投稿的完整工作流



实证研究智能体技能大全封面图

🚀 New here? Open the Skill Search → to filter all 1,096 skills by method, stage, language, and license. The 5-minute tour (make quickstart) prints the same picture in your terminal.

🇨🇳 中文用户从本文件开始(流水线速览 + 76 行总表),每个合集的完整描述见 docs/CONTENT_ZH.md。📖 English readers: see README-en.md.


信任面 · Trust surface (rigor stats)

Rigor laneCountWhere
Numeric benchmark tasks — gold values recomputed from real data each run17benchmark/
Behavioral eval scenarios / rubric items37 / 183eval-harness/

Full trust overview: docs/TRUST.md · docs/RIGOR_COVERAGE.md


中文文档结构

中文内容分两级维护,各司其职:

  • 本文件(README.md,GitHub 默认入口):banner、badges、信任面、9 阶段流水线速览、76 行合集总表。
  • docs/CONTENT_ZH.md(扩展正文):每个合集的完整描述(#skill-NN 锚点)、按用途分组、精确数字、2 分钟验证、三层信任、旗舰流水线详解、贡献与引用。总表行内的 → 直接跳到对应锚点。
  • 其他语言:README-en.md · README-zh-TW.md · README-ja.md · README-ko.md

[!NOTE] 维护规则: 改合集总表 → 本文件与 CONTENT_ZH.md 的锚点表两处同步;改合集详情 / 分组 / 数字 → 只改 docs/CONTENT_ZH.md。统计数字(合集数 / skill 数)以 catalog/skills.json 为准,由 make validate 的 readme-stats 检查器守护。

贡献者(Contributors): 提交前请在本地跑通完整门禁 make check(catalog 校验 + 链接 + 单元测试 + eval-harness + benchmark)。详见 CONTRIBUTING.md。

旧版归档: README-zh-CN.md 已弃用,仅作向后兼容的重定向占位。


🚀 从一个 idea 到一篇论文:社科实证研究 · 端到端流水线(全自动、可介入)

AERS 不只是 76 个散装 skill —— 它能陪你走完一篇论文。 从模糊 idea → 选题精炼 → 文献综述 → 数据获取 → 识别策略 → 估计建模 → 稳健性审计 → 出版级表格 / 图形 → 写作与同行评审 → 降 AIGC → 投稿。端到端、全自动、每一步都可被人介入(中间任何一步你都可以接过去手工改方法、补变量、加稳健性,再让流水线自动接上跑)。

9 阶段流水线 · 每一步都覆盖到具体 skill

#阶段关键 skills(点合集名进目录,→ 进完整说明)
1️⃣选题精炼 — Agent 把模糊想法收紧成"可证伪 + 可执行"的研究问题· 25 Diverga · 33 claude-scholar · 05 research-superpower · 11 compound-science
2️⃣文献综述 — 检索 · 筛选 · PRISMA 流程 · 批判性阅读 · 主题分析· 36 literature-review-skill · 24 academic-research-skills · 59 openalex-skill · 68 research-productivity-skills · 53 thematic-analysis
3️⃣数据获取 — 公开数据库 · API · 网页抓取 · 数据清洗· 33 claude-scholar · 68 research-productivity-skills · 32 stata-skill · 57 edgartools
4️⃣识别策略 — DiD / RD / IV / SCM / DML / matching 全覆盖· ⭐ 00 StatsPAI 🔥 · 10 causal-inference-mixtape · 13 MixtapeTools · 51 CausalPy · 63 scientific-agent-skills
5️⃣估计建模 — Python / Stata / R 三栈,900+ 估计器· ⭐ 00.1 Full Empirical · Python · ⭐ 00.2 Full Empirical · Stata · ⭐ 00.3 Full Empirical · R · 40 pyfixest · 39 marginaleffects · 09 awesome-econ-ai
6️⃣稳健性审计 — 复现包检查 · Honest-DiD · R&R 模拟· 41 sewage-econometrics-check · ⭐ 50 AER-skills · 21 AI-research-feedback
7️⃣表格 & 图形 — 期刊出版级排版 · LaTeX 嵌入· ⭐ 00 StatsPAI · 07 AI-Research-SKILLs · 33 claude-scholar · 08 latex-document-skill
8️⃣写作 & 同行评审 — LaTeX / Quarto · 仿审稿人 · 校对· 06 stats-paper-writing · 04 scientific-writer · 22 christopherkenny-skills · 38 academic-proofreader · 56 econ-writing-skill · 16 clo-author
9️⃣降 AIGC & 投稿 — 知网 / 万方 / Turnitin / 23 类 AI 痕迹模式· ⭐ 48 de-AIGC-skills 🇨🇳🇬🇧 · 44 humanizer_academic · 45 deslop · 46 stop-slop · 47 avoid-ai-writing · 49 humanize-chinese

🎼 元编排:⭐ 69 Paper-WorkFlow —— 一键串起来

Paper-WorkFlow 是 AERS 的"指挥棒",它把上面 9 个阶段的 skill 串成 一条按键即运行的端到端流水线。 你在 IDE 入口给它一句自然语言:

"开一个新论文项目:空气污染与中国劳动力市场,CS 设计 + 省级面板"

它会自动按顺序调:

  1. ⭐ 00 StatsPAI → sp.csdid(...) 给出 CS-DID 估计草案 + 写出估计方程与识别假设
  2. 33 claude-scholar → 抓变量定义 / 数据源候选 / 相关文献
  3. ⭐ 00 StatsPAI → 真跑 sp.feols(...) + sp.honest_did(...)
  4. 41 sewage-econometrics-check → 10 项复现包审计 + 稳健性体检
  5. ⭐ 00 StatsPAI + 07 AI-Research-SKILLs → 出 Table 1–5 + 期刊级图
  6. 38 academic-proofreader → 通读 + §comment 标"审稿人会挑刺的位置"
  7. 56 econ-writing-skill 起草初稿 + ⭐ 48 de-AIGC-skills 🇨🇳🇬🇧 + 45 deslop 过知网 / Turnitin

任何阶段你都可以手动介入 —— 上一阶段的产物全部落盘(产物-幂等 pipeline),你接过去改方法、补控制、加稳健性,再让流水线自动接下去跑。这就是"全自动 + 可介入"。

🏆 7 个 Stanford REAP × CoPaper.AI 自研 skill —— 是整个流水线的主干

⭐ Skill在流水线里的角色
00 StatsPAI 🔥因果引擎:900+ 函数,sp.causal(...) 一行跑闭环(DID / RD / IV / SCM / DML / matching)
00.1 Full Empirical · Python 📘显式 Python 栈(pandas / statsmodels / linearmodels / pyfixest)
00.2 Full Empirical · Stata 📊显式 Stata 栈(reghdfe / ivreg2 / csdid / sdid / rdrobust)
00.3 Full Empirical · R 📗显式 R 栈(tidyverse / fixest / did / HonestDiD)+ Quarto 渲染
48 de-AIGC-skills 🇨🇳🇬🇧中英双语学术降 AIGC(Turnitin AI / GPTZero / 知网 / 万方)
50 AER-skills 📕Top-5 经济学投稿套件:识别 → 稳健性 → R&R
69 Paper-WorkFlow 🧭元编排器,把上面 9 个阶段串成一键流水线

为什么挑这 7 个?因为它们的行为都被基准钉死了 —— 不是营销口径,是对着已知答案反复跑过验证过的(17 项数值 benchmark + 37 项行为评测 ↗)。

看到这里 —— 完整 76 行合集目录

↴ 直跳到下方 76 行总表(每个合集带 #skill-NN 锚点)。如果你更关心"这些 skill 怎么用"而不是"有哪些 skill",看 📘 中文唯一权威正文 里的「按用途分组」与「旗舰流水线」两节。


🧰 76 个核心 Skills 合集一览(00 → 72,编号连续无空缺)

打开仓库 → 看见整座库。 全部 76 个合集 · 1,096 个 skill,每一个都已 vendor 进本仓库,由 catalog/skills.json 跟踪。⭐ = Stanford REAP × CoPaper.AI 团队自研的 skill;其余为精选、经安全审计的社区作品。

主题图例 — 🚀 全流程与编排器 · 🎯 因果推断与计量经济学 · 📚 文献与研究设计 · ✍️ 写作 / 编辑 / 去 AIGC · 📑 引用 / 复现 / 同行评审 · 🛠️ 数据 / 工具 / 基础设施

点击【→】 跳转到 docs/CONTENT_ZH.md 中该合集的完整描述;点击合集名 直接打开其目录。

#合集一句话详情
⭐ 00StatsPAI 🔥因果引擎 · Agent-native Python DSL:sp.causal(...) 一行跑闭环(DID/RD/IV/SCM/DML,900+ 函数)→
⭐ 00.1Full Empirical · Python 📘显式栈:pandas · statsmodels · linearmodels · pyfixest→
⭐ 00.2Full Empirical · Stata 📊reghdfe · ivreg2 · csdid · sdid · rdrobust 复现包→
⭐ 00.3Full Empirical · R 📗tidyverse · fixest · did · HonestDiD + Quarto 渲染→
01academic-paper-skills大纲 → 手稿写作 + 7 维审稿人模拟→
02research-skills医学影像综述、提案、论文转幻灯片→
03scientific-skills假设生成 + 28 个科学数据库→
04scientific-writer引用管理 + 科学写作→
05research-superpower系统化检索、筛选与引文溯源→
06stats-paper-writing端到端 LaTeX 统计论文写作→
07AI-Research-SKILLs发表级 ML 图表、LaTeX、引文核验→
08latex-document-skill创建 / 编译任意 LaTeX 文档为 PDF→
09awesome-econ-aiPython 面板数据分析(linearmodels)→
10causal-inference-mixtapeDID / IV / RDD / SCM 模板(Cunningham)→
11compound-science面向定量社会科学的贝叶斯估计→
12claude-code-my-workflow提交 → PR → 合并的研究工作流(Emory)→
13MixtapeToolsCunningham 的因果推断工具集与讲义→
14research-starterR 中的 IV / DiD / RDD,含完整诊断→
15social-science-researchR 或 Python 端到端数据分析→
16clo-author多代理数据分析(R / Stata / Python)→
17DAAF安全意识代理框架(32 条 deny rule)→
18stata-accounting来自 126 篇 JAR 论文的实测 Stata 范式→
19vera-economic-intelligence经济情报 / 政策研究情报工作流→
20python-econ-skillDSGE / HANK 与定量经济计算→
21AI-research-feedback用 AI 同行评审生成结构化反馈→
22christopherkenny-skills面向 Quarto(.qmd)的 APSA 风格检查器→
23baygent带护栏的 PyMC / Arviz 贝叶斯工作流→
24academic-research-skills5 审稿人多视角论文评审→
25Diverga研究问题精炼器(抗模式坍缩)→
26scholar统计算法设计与文档→
27my_claude_skills经济学摘要写作指南→
28paper-replicate-agent论文复现代理演示→
29project20XXy可复现手稿 + notebook 项目→
30zirui-song-claude-skillsZirui Song 的研究辅助 Claude 技能集→
31claude-code-skillsPython 面板数据分析→
32stata-skill高性能 Stata C/C++ 插件→
33claude-scholar研究全生命周期:选题 → 综述 → 实验 → 审稿回复→
34research-companion头脑风暴、评估并决策研究方向→
35academic-writing-skills面向投稿场所的工业 AI 文献研究→
36literature-review-skill完整文献综述工作流(中文)→
37IlanStrauss-ai-skillsIlan Strauss 经济学研究 AI 工作流→
38academic-proofreader学术校对→
39marginaleffects预测、斜率与比较(R / Python)→
40pyfixestPython 中的快速固定效应估计→
41sewage-econometrics-check10 项复现包审计→
42ARIS自主「research-in-sleep」代理,端到端→
43research-plugins478 个研究插件:数据可视化、领域、基础设施→
44humanizer_academic为医学/学术手稿去 AI 味(23 类模式)→
45deslop去除 AI 写作痕迹(5 维评分)→
46stop-slop三层 AI 痕迹检测与改写→
47avoid-ai-writing审计 → 改写 → 二次审计 AI 味(留痕)→
⭐ 48de-AIGC-skills 🇨🇳🇬🇧中英双语学术降 AIGC(Turnitin AI / GPTZero / 知网 / 万方)→
49humanize-chinese检测并人性化 AI 生成的中文文本→
⭐ 50AER-skills 📕Top-5 经济学投稿套件:识别 → 稳健性 → R&R→
51CausalPy贝叶斯准实验(PyMC Labs)→
52slr-prisma系统文献综述,PRISMA 2020→
53thematic-analysisBraun & Clarke 六阶段定性主题分析→
54open-science-skills引用一致性、DOI 与论据支撑审计→
55r-skillsR 中用 brms 做贝叶斯推断→
56econ-writing-skill综合 50+ 顶级指南的经济学写作→
57edgartools查询与分析 SEC 文件→
58econstack政策简报(UK GES / AU Treasury)→
59openalex-skill通过 OpenAlex 查询 2.4 亿+ 学术作品→
60superpapers综合性实证研究支持套件→
61research-methods与预注册匹配的验证性检验→
62citation-checker对照 CrossRef / S2 / OpenAlex 核验引用→
63scientific-agent-skillsDoWhy 识别–估计–反驳框架→
64mcp-stata20 个 Stata 因果推断与复现 skill→
65game-theory-paper-writer生成并压力测试博弈论论文→
66empirical-research-skills面向大型面板的 R 性能优化→
67econfin-workflow-toolkit中国公司金融实证工作流,从提案到论文→
68research-productivity-skills论文检索、SSRN、DOI 查询、下载→
⭐ 69Paper-WorkFlow 🧭元编排器,串起整个社会科学论文流水线→
70ssci-polish ✍️SSCI / SCI 英文论文语言润色(语法、可读性、学术语气)→
⭐ 71lit-review-agent-tools 🔍文献综述工具选型 + 一键安装运行(MinerU / PaperQA2 / ASReview / STORM / MCP 服务器)→
⭐ 72Kaggle Research 🧪通过官方 CLI 安全检索 Kaggle 资源、限界下载公开数据并保留审计证据→

想看更详细的描述(主题分类、字段、统计)? 见 docs/CONTENT_ZH.md 中标注 #skill-NN 锚点的同一张表 —— 它是每个合集的完整描述所在的扩展正文。


AI 是放大器,不是替代品。它替你做最耗时的"搬砖",你保留最核心的"判断"。


CoPaper.AI Stanford REAP

Stanford REAP × CoPaper.AI · 实证研究 AI 工具的学术工业级产品


扫码访问 copaper.ai
扫码访问 copaper.ai
CoPaper.AI 公众号
关注公众号「CoPaper.AI」

内置 20 个方法论 skill · 20 分钟完成实证论文 · 自研 StatsPAI(900+ 函数 / MIT 开源)

内容与创作Agent / MCP / Skill 创作

中风险

  • 来源需自行核对维护者身份。
  • 包含脚本或命令调用,安装前请复核。
  • 未检测到明显外部权限要求。
  • 未检测到高风险命令。
  • 扫描发现:2 条。

Codex — Git Clone 安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 克隆仓库:git clone https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills.git
  3. 将 "skills/01-lishix520-academic-paper-skills/composer" 文件夹复制到 Codex 的 skills 目录中。
  4. 重启 Codex 让新的 skill 生效。

Codex — 手动复制安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 从源仓库下载 SKILL.md 及相关文件。
  3. 在 Codex 的 skills 目录中创建新文件夹。
  4. 将所有 skill 文件复制到新文件夹中。
  5. 重启 Codex 让新的 skill 生效。

Claude Code — Git Clone 安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 克隆仓库:git clone https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills.git
  3. 将 "skills/01-lishix520-academic-paper-skills/composer" 文件夹复制到 Claude Code 的 skills 目录中。
  4. 重启 Claude Code 让新的 skill 生效。

Claude Code — 手动复制安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 从源仓库下载 SKILL.md 及相关文件。
  3. 在 Claude Code 的 skills 目录中创建新文件夹。
  4. 将所有 skill 文件复制到新文件夹中。
  5. 重启 Claude Code 让新的 skill 生效。

Cursor — Git Clone 安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 克隆仓库:git clone https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills.git
  3. 将 "skills/01-lishix520-academic-paper-skills/composer" 文件夹复制到 Cursor 的 skills 目录中。
  4. 重启 Cursor 让新的 skill 生效。

Cursor — 手动复制安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 从源仓库下载 SKILL.md 及相关文件。
  3. 在 Cursor 的 skills 目录中创建新文件夹。
  4. 将所有 skill 文件复制到新文件夹中。
  5. 重启 Cursor 让新的 skill 生效。

GitHub Copilot — Git Clone 安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 克隆仓库:git clone https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills.git
  3. 将 "skills/01-lishix520-academic-paper-skills/composer" 文件夹复制到 GitHub Copilot 的 skills 目录中。
  4. 重启 GitHub Copilot 让新的 skill 生效。

GitHub Copilot — 手动复制安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 从源仓库下载 SKILL.md 及相关文件。
  3. 在 GitHub Copilot 的 skills 目录中创建新文件夹。
  4. 将所有 skill 文件复制到新文件夹中。
  5. 重启 GitHub Copilot 让新的 skill 生效。

Windsurf — Git Clone 安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 克隆仓库:git clone https://github.com/brycewang-stanford/Auto-Empirical-Research-Skills.git
  3. 将 "skills/01-lishix520-academic-paper-skills/composer" 文件夹复制到 Windsurf 的 skills 目录中。
  4. 重启 Windsurf 让新的 skill 生效。

Windsurf — 手动复制安装

  1. 安装前请先查看来源仓库和风险报告。
  2. 从源仓库下载 SKILL.md 及相关文件。
  3. 在 Windsurf 的 skills 目录中创建新文件夹。
  4. 将所有 skill 文件复制到新文件夹中。
  5. 重启 Windsurf 让新的 skill 生效。
查看 SKILL.md 原文
name: academic-paper-composer
description: 'Systematic writing framework for philosophy and interdisciplinary academic papers from optimized outline to submission-ready manuscript. Use when users want to: (1) write a paper from a detailed outline, (2) ensure quality control during writing, (3) maintain consistency across chapters, (4) prepare a submission-ready manuscript, or (5) systematically execute a planned paper. Triggered by phrases like ''write the paper from this outline,'' ''compose the full manuscript,'' ''execute the outline,'' or when users have completed strategic planning (academic-paper-strategist skill) and are ready to write. Takes optimized outline as input; outputs complete manuscript with iterative quality checks.'

Academic Paper Composer

Overview

This skill provides a systematic framework for writing academic papers from optimized outline to submission-ready manuscript. It implements iterative quality control at both chapter-level and paper-level, ensuring consistent high quality throughout the writing process.

Input: Detailed, optimized paper outline (from academic-paper-strategist or equivalent)

Output: Complete, submission-ready manuscript with quality validation reports

Prerequisite: Use academic-paper-strategist skill first to create optimized outline (or provide equivalent detailed outline)


When to Use This Skill

Use academic-paper-composer when you need to:

Writing Stage:

  • Execute a detailed paper outline systematically
  • Write a complete academic paper chapter-by-chapter
  • Maintain quality control during writing process
  • Ensure consistency across all sections

Quality Assurance Stage:

  • Validate each chapter before proceeding
  • Check cross-chapter coherence
  • Perform final quality assessment
  • Verify submission readiness

Submission Preparation Stage:

  • Prepare manuscript for platform submission
  • Complete platform-specific checklists
  • Generate final submission package

Triggers:

  • "Write the paper from this outline"
  • "Compose the full manuscript"
  • "Execute the outline systematically"
  • "I have an outline, help me write the paper"
  • After completing academic-paper-strategist skill

Workflow Overview

Phase 4: SYSTEMATIC WRITING (Chapter-by-Chapter + Quality Gates)
    ↓
Phase 5: QUALITY CONTROL (Final Validation + Submission Prep)
    ↓
Output: Submission-Ready Manuscript + Quality Reports

Quality Gates: After each chapter + final paper evaluation


Required Input

Before using this skill, you must provide:

1. Optimized Detailed Outline

Required structure:

## Abstract (250-300 words)
- [Key points to cover]

## 1. Introduction (1,500 words)
### 1.1 Opening Puzzle (400 words)
- [Content guidance]
### 1.2 Literature Review (600 words)
- [Theories to discuss]
### 1.3 This Paper's Contribution (500 words)
- [Specific claims]

## 2. [Main Chapter Title] (1,200 words)
### 2.1 [Section] (400 words)
- [Argument structure]
- [Key citations]
...

[Complete structure with word counts and content guidance]

Quality check: Outline should specify:

  • ✓ Chapter titles and word counts
  • ✓ Subsection structure (to 3rd level)
  • ✓ Content guidance for each section
  • ✓ Key citations to include
  • ✓ Argument structure notes

If outline lacks these, consider using academic-paper-strategist first.

2. Platform Writing Standards Guide

From academic-paper-strategist Phase 1, or equivalent document specifying:

  • Platform style patterns (voice, terminology, citation format)
  • Structural conventions
  • Example papers for reference

3. Literature Base (Optional but Recommended)

  • List of core papers to cite
  • Research gap analysis
  • Key concepts to emphasize

Phase 4: Systematic Writing

Goal

Write complete manuscript chapter-by-chapter with iterative quality control.

Workflow

Step 4.1: Writing Environment Setup

Before writing, I will:

  1. Verify outline completeness:

    • All chapters specified with word counts
    • Content guidance provided
    • Argument structure clear
  2. Load reference documents:

    • Platform writing standards
    • Section writing guides (references/section_guides.md)
    • Quality standards (references/writing_standards.md)
  3. Create writing tracker:

    # Writing Progress Tracker
    - [ ] Abstract (250 words)
    - [ ] Chapter 1: Introduction (1,500 words)
    - [ ] Chapter 2: [Title] (1,200 words)
    ...
    - [ ] Conclusion (1,000 words)
    - [ ] References
    

Decision Point 1: Confirm outline and standards loaded, ready to begin writing.

Step 4.2: Chapter-by-Chapter Writing

For each chapter, I will follow this sequence:

A. Pre-Writing Review

Before writing chapter N:

  1. Review outline specification for this chapter:

    • Word count target
    • Subsection structure
    • Content guidance
    • Key arguments/citations
  2. Review previous chapter (if N>1):

    • Last paragraph of chapter N-1
    • Key concepts introduced
    • Promises to fulfill
  3. Check section guide:

    • Load appropriate template from references/section_guides.md
    • Review quality markers for this section type
B. Writing Execution

I will write the chapter following:

Content principles:

  • Follow outline exactly: Respect structure and word counts
  • Include specified citations: Use literature from outline
  • Maintain platform style: Match voice and terminology
  • Use section templates: Follow appropriate guide from section_guides.md

Quality targets (pre-emptive):

  • Argument quality: Clear thesis, justified premises
  • Citation quality: All claims supported, proper format
  • Clarity: Precise prose, terms defined, good transitions
  • Structure: Logical flow, proper proportions
  • Style conformity: Match platform conventions

Output: Complete chapter draft

C. Post-Writing Evaluation

After completing chapter draft, I will:

  1. Create evaluation document:

    python scripts/chapter_quality_check.py
    # Option 1: Create template for this chapter
    
  2. Perform 5-dimension assessment:

    • Argument Quality (1-4): Thesis clear? Premises justified? Objections addressed?
    • Citation Quality (1-4): All claims cited? Format consistent? Key literature included?
    • Clarity & Readability (1-4): Prose clear? Terms defined? Transitions smooth?
    • Structure & Flow (1-4): Logical progression? Proper proportions? Follows outline?
    • Platform Conformity (1-4): Style match? Voice consistent? Format correct?
  3. Generate quality report:

    python scripts/chapter_quality_check.py
    # Option 2: Generate report from evaluation
    

    This produces:

    • Total score (X/20)
    • Pass/fail (threshold: ≥16/20)
    • Weak dimensions identified
    • Specific revision recommendations

Quality Gate 4A (After Each Chapter):

  • ✓ Score ≥16/20 (80%)
  • ✓ All dimensions ≥3/4 (or revisions implemented)
  • ✓ Word count within ±10% of target

If Failed: Implement revisions before proceeding to next chapter

D. Iteration (If Needed)

If chapter scores <16/20:

  1. Identify weak dimension(s): Which scored <3/4?

  2. Implement targeted revisions:

    • Argument quality issue → Add justifications, address objections
    • Citation quality issue → Add supporting citations, fix format
    • Clarity issue → Simplify prose, add definitions, improve transitions
    • Structure issue → Reorganize paragraphs, adjust proportions
    • Style issue → Adjust voice, terminology, format
  3. Re-evaluate: Generate new quality report

  4. Repeat until passing (typically 1-2 iterations)

Important: Do not proceed to next chapter until current chapter passes quality gate.

E. Chapter Completion

Once chapter passes quality gate:

  1. Mark chapter complete in writing tracker
  2. Save chapter with quality report
  3. Note key concepts introduced (for coherence check later)
  4. Preview next chapter requirements

Decision Point 2 (After Major Chapters): After completing each main body chapter, I will:

  • Present completed chapter summary
  • Show quality score
  • Ask: Proceed to next chapter or revise further?

Step 4.3: Writing Sequence

Recommended order:

  1. Introduction (write first)

    • Establishes thesis and roadmap
    • Use introduction guide from section_guides.md
    • Quality check: Abstract promises, literature coverage, clear contribution
  2. Main Body Chapters (in outline order)

    • Follow outline sequence
    • Quality check after each
    • Build on previous chapters
  3. Conclusion (write after main body)

    • Synthesizes findings
    • Addresses introduction promises
    • Use conclusion guide from section_guides.md
  4. Abstract (write last)

    • Summarizes complete paper
    • Use abstract guide from section_guides.md
    • Quality check: Standalone, accurate, compelling
  5. References (compile throughout)

    • Format according to platform standards
    • Verify all citations present

Step 4.4: Cross-Chapter Coherence Check

After all chapters written, before final evaluation:

  1. Terminology consistency:

    • Extract key terms from each chapter
    • Verify consistent usage throughout
    • Check definitions consistent
  2. Argument flow:

    • Verify chapter N+1 builds on chapter N
    • Check roadmap (intro) matches execution
    • Ensure conclusion addresses introduction promises
  3. Citation patterns:

    • Check for uneven citation distribution
    • Verify key works cited where relevant
    • Ensure bibliography complete

Output: Cross-chapter coherence report identifying any inconsistencies


Phase 5: Quality Control

Goal

Perform comprehensive final evaluation and prepare submission-ready manuscript.

Workflow

Step 5.1: Content Completeness Check

Using structured checklist, verify:

Structural Completeness:

  • Abstract (250-300 words)
  • Introduction with all required elements
  • All outlined main chapters present
  • Conclusion with all required elements
  • References section formatted correctly

Content Completeness:

  • All introduction promises fulfilled
  • All claims supported by evidence or argument
  • All technical terms defined
  • All objections addressed
  • All limitations acknowledged

Citation Completeness:

  • Every citation in text has bibliography entry
  • Every bibliography entry cited in text
  • Citation format consistent throughout
  • All citations include necessary information

Format Completeness:

  • Title page (if required)
  • Section numbering consistent
  • Heading hierarchy logical
  • Figure/table captions (if applicable)

Output: Completeness checklist report

Step 5.2: Final 7-Dimension Evaluation

I will perform comprehensive evaluation using:

python scripts/final_evaluation.py
# Create evaluation template

7 Dimensions (10 points each, 70 total):

  1. Overall Argument Quality (1-10)

    • Thesis clarity throughout
    • Chapter integration
    • Logical completeness
    • Objections addressed
  2. Literature Integration (1-10)

    • Citation count (40-60 typical for philosophy)
    • Key literature covered
    • Citations well-integrated
    • Critical engagement present
  3. Clarity & Accessibility (1-10)

    • Prose clarity
    • Complex ideas explained
    • Appropriate for audience
    • Technical terms defined
  4. Originality & Contribution (1-10)

    • Clear original contribution
    • Advance over literature
    • Significance established
    • Innovation present
  5. Methodological Rigor (1-10)

    • Method explicit and justified
    • Consistently applied
    • Appropriate for question
    • Limitations acknowledged
  6. Structure & Organization (1-10)

    • Logical flow
    • Optimal organization
    • Proportions balanced
    • Transitions seamless
  7. Platform & Style Conformity (1-10)

    • Style matches platform
    • Format correct
    • Voice consistent
    • Citation format perfect

Scoring Process:

  1. Evaluate each dimension (1-10)
  2. Provide detailed notes for each
  3. Complete completeness checklist
  4. List any specific issues

Generate report:

python scripts/final_evaluation.py
# Generate final report

This produces:

  • Total score (X/70)
  • Pass/fail (threshold: ≥56/70)
  • Weak dimensions identified
  • Completeness assessment
  • Prioritized revision recommendations
  • Submission readiness decision

Quality Gate 5 (Final):

  • ✓ Score ≥56/70 (80%)
  • ✓ All completeness checklist items complete
  • ✓ All high-priority issues addressed

If Failed: Implement revisions and re-evaluate

Step 5.3: Revision Implementation (If Needed)

If final score <56/70 or completeness incomplete:

  1. Prioritize revisions:

    • HIGH priority: All issues affecting score or completeness
    • MEDIUM priority: Issues improving quality
    • LOW priority: Optional enhancements
  2. Implement systematically:

    • Address high-priority first
    • Document changes
    • Maintain style consistency
  3. Re-evaluate:

    • Generate new final report
    • Verify score ≥56/70
    • Check completeness 100%
  4. Iterate until passing

Decision Point 3: After final evaluation, I will:

  • Present final score and assessment
  • Show submission readiness status
  • Recommend: Submit immediately / Implement optional improvements / Required revisions
  • Ask: Proceed with submission or implement further improvements?

Step 5.4: Submission Package Preparation

Once final evaluation passes:

  1. Platform-specific checklist:

    • PhilArchive/PhilPapers:

      • PDF format
      • Abstract <500 words
      • Metadata (title, keywords, classification)
      • Author information complete
    • arXiv:

      • LaTeX or PDF format
      • Abstract <1920 characters
      • Category selection correct
      • No font embedding issues
    • PhilSci-Archive:

      • PDF format
      • Subject classification
      • Keywords (3-5)
      • No copyright issues
  2. Generate final outputs:

    • Formatted manuscript (PDF or LaTeX)
    • Abstract (separate file if needed)
    • Metadata file
    • Cover letter (if applicable)
  3. Pre-submission verification:

    • Re-read complete paper
    • Check all formatting
    • Verify all links/citations work
    • Proofread for typos

Output: Complete submission package ready for platform upload


Complete Output Package

Upon completion of both phases, you receive:

Quality Reports

  1. Chapter Quality Reports (one per chapter)

    • 5-dimension scores
    • Pass/fail status
    • Revision recommendations implemented
  2. Cross-Chapter Coherence Report

    • Terminology consistency check
    • Argument flow verification
    • Citation pattern analysis
  3. Final Evaluation Report ⭐ Key Document

    • 7-dimension comprehensive assessment
    • Completeness checklist (100%)
    • Submission readiness decision
    • Platform-specific checklist

Manuscript Files

  1. Complete Manuscript ⭐ Main Deliverable

    • All chapters integrated
    • Properly formatted
    • Citations complete
    • Submission-ready
  2. Abstract (separate file)

    • Standalone 250-300 words
    • Platform-formatted
  3. Metadata Document

    • Title, keywords, classification
    • Author information
    • Platform-specific requirements

Supporting Documentation

  1. Writing Progress Tracker

    • All chapters completed
    • Quality scores logged
    • Revision history
  2. Citation List

    • All references used
    • Formatted for platform
    • Verified complete

Quality Assurance System

Quality Standards Reference

For detailed evaluation criteria, all standards are defined in:

references/writing_standards.md

This document provides:

  • Chapter-level quality standards (5 dimensions)
  • Final quality standards (7 dimensions)
  • Section-specific quality criteria
  • Scoring rubrics
  • Issue identification and fixes

Section Writing Guides

For guidance on writing each section type:

references/section_guides.md

This provides:

  • Abstract writing template and checklist
  • Introduction structure (6 subsections)
  • Main body chapter templates (4 types)
  • Conclusion structure
  • Transition strategies
  • Platform-specific style guidance
  • Common mistakes and fixes

Evaluation Scripts

Two Python scripts support quality validation:

1. Chapter Quality Check

python scripts/chapter_quality_check.py

Function: Evaluates individual chapters against 5-dimension standards

  • Creates evaluation templates
  • Calculates scores and pass/fail
  • Identifies weak dimensions
  • Generates revision recommendations
  • Supports multi-chapter comparison

When to Use: After writing each chapter (Step 4.2.C)

2. Final Paper Evaluation

python scripts/final_evaluation.py

Function: Evaluates complete paper against 7-dimension standards

  • Comprehensive quality assessment
  • Completeness checklist validation
  • Submission readiness determination
  • Platform-specific preparation
  • Generates final report with recommendations

When to Use: After all chapters complete (Step 5.2)


Decision Points (Interactive)

This skill has 3 key decision points where I pause for your input:

Decision Point 1: Writing Commencement (Step 4.1)

I provide: Loaded outline, standards, and writing plan You confirm: Ready to begin writing / Need adjustments

Decision Point 2: Chapter Completion (Step 4.2.E)

I provide: Completed chapter with quality score You decide: Proceed to next / Revise further / Adjust approach

Frequency: After each main body chapter (3-5 times typically)

Decision Point 3: Final Submission (Step 5.3)

I provide: Final evaluation report with score and readiness assessment You decide: Submit immediately / Implement optional improvements / Additional revisions


Example Usage

User Request

"I have the optimized outline from the strategist skill. Please write the complete paper on self-continuity during sleep."

Skill Response

Phase 4: Systematic Writing

  1. Setup: Outline loaded (6 chapters, 9,600 words total). Platform: PhilArchive. Standards loaded. ✓

  2. Chapter 1 - Introduction (1,500 words):

    • Writing... [follows outline structure]
    • Self-evaluation: Argument 4/4, Citation 3/4, Clarity 4/4, Structure 3/4, Style 4/4
    • Score: 18/20 ✓ PASS
    • Decision Point 2: Proceed to Chapter 2?
  3. Chapter 2 - Narrative Compression Framework (1,200 words):

    • Writing... [develops theoretical framework]
    • Self-evaluation: All dimensions 3-4
    • Score: 17/20 ✓ PASS
    • Decision Point 2: Proceed to Chapter 3?
  4. Chapter 3-5: [Similar process for each chapter]

  5. Chapter 6 - Conclusion (1,000 words):

    • Writing... [synthesizes findings]
    • Score: 18/20 ✓ PASS
  6. Abstract (280 words):

    • Writing... [summarizes complete paper]
    • Score: 19/20 ✓ PASS
  7. Cross-Chapter Coherence: Terminology consistent ✓, Argument flow verified ✓, Citations complete ✓

Phase 5: Quality Control

  1. Completeness Check: All items complete ✓ (18/18)

  2. Final Evaluation:

    • Overall Argument: 9/10
    • Literature Integration: 8/10 (48 citations)
    • Clarity: 9/10
    • Originality: 8/10
    • Methodology: 8/10
    • Structure: 9/10
    • Platform Fit: 9/10
    • Total: 60/70 (85.7%) ✓ PASS
  3. Decision Point 3: Paper scored 60/70 (Good quality, ready for submission). Submit now or implement optional improvements?

  4. Submission Package: PDF generated, metadata prepared, PhilArchive checklist complete ✓

Output: Submission-ready manuscript + comprehensive quality reports


Tips for Best Results

Provide Complete Outline

  • The more detailed your outline, the better the writing quality
  • Include content guidance for each section
  • Specify key citations and argument structure
  • Indicate expected word counts

Trust the Iterative Process

  • Chapter-level quality gates ensure each section is solid
  • Don't skip quality checks (they prevent cascading problems)
  • Revisions at chapter-level are easier than whole-paper revisions

Engage at Decision Points

  • Your input at decision points shapes the final manuscript
  • Feel free to request additional revisions even if chapter passes
  • Provide feedback on style or emphasis

Use Quality Reports

  • Chapter reports show specific weaknesses
  • Final report provides objective submission readiness assessment
  • Use reports to track improvement across chapters

Leverage Section Guides

  • I reference section_guides.md for each section type
  • You can review these guides directly if you want to understand the approach
  • Guides include templates, examples, and common mistakes

Integration with Academic-Paper-Strategist

This skill is designed to work seamlessly with academic-paper-strategist:

Ideal workflow:

  1. Use academic-paper-strategist to:

    • Identify optimal platform
    • Conduct literature search
    • Identify research gaps
    • Assess originality
    • Generate optimized detailed outline
  2. Use academic-paper-composer (this skill) to:

    • Execute the outline systematically
    • Maintain quality control during writing
    • Produce submission-ready manuscript

Can be used standalone: If you already have a detailed outline from another source, you can use this skill directly (skip strategist).


Limitations and Notes

  • Requires detailed outline: Vague outlines produce lower-quality output; specificity is key
  • Iterative process takes time: Quality writing with validation requires patience; typical timeline: 1-2 days for 10,000-word paper
  • Quality checks are systematic, not perfect: Final human review recommended before submission
  • Platform-specific formatting: I adapt to platform standards, but you should verify final format
  • Complementary to strategist skill: Best results come from using both skills in sequence

Common Issues and Solutions

Issue 1: Chapter Fails Quality Gate

Symptom: Score <16/20 after writing chapter

Solution:

  1. Review weak dimension(s) from report
  2. Implement specific recommendations
  3. Re-evaluate chapter
  4. Typical fix time: 30-60 minutes

Prevention: Follow section guides closely during initial writing

Issue 2: Inconsistent Style Across Chapters

Symptom: Some chapters feel different in tone or voice

Solution:

  1. Run cross-chapter coherence check (Step 4.4)
  2. Identify inconsistent terminology or voice
  3. Revise to match dominant style
  4. Re-run check to verify

Prevention: Reference platform standards before writing each chapter

Issue 3: Low Final Score (<56/70)

Symptom: Paper fails final quality gate

Solution:

  1. Identify weak dimensions from final report
  2. Focus on dimensions scoring <7/10
  3. Implement high-priority revisions systematically
  4. Re-evaluate after revisions

Common causes: Insufficient literature integration, unclear contribution, poor coherence

Issue 4: Completeness Checklist Incomplete

Symptom: Missing required elements

Solution:

  1. Review which category has incomplete items
  2. Add missing elements (e.g., missing objections section, incomplete references)
  3. Re-run completeness check

Prevention: Use writing tracker throughout; check outline completeness before starting


Platform-Specific Notes

PhilArchive / PhilPapers

  • Style: First-person acceptable ("I argue")
  • Length: 5,000-12,000 words typical
  • Citations: APA or Chicago author-year
  • Quality focus: Philosophical rigor, argument clarity

arXiv (Philosophy-adjacent)

  • Style: More formal, passive voice common
  • Length: Varies widely (3,000-20,000)
  • Citations: Varies by subcategory
  • Quality focus: Interdisciplinary clarity, technical precision

PhilSci-Archive

  • Style: Bridges philosophical and scientific
  • Length: 6,000-15,000 words typical
  • Citations: Author-year typical
  • Quality focus: Integration of philosophy + science

Summary

academic-paper-composer transforms an optimized outline into a submission-ready manuscript through:

  1. Systematic Writing (Phase 4): Chapter-by-chapter execution with 5-dimension quality checks after each (≥16/20 threshold)
  2. Quality Control (Phase 5): Final 7-dimension assessment (≥56/70 threshold) + completeness validation + submission preparation

Quality Assurance: Iterative evaluation at chapter and paper levels ensures consistent quality throughout.

Output: Submission-ready manuscript with comprehensive quality reports documenting systematic validation.

Estimated Time: 1-2 days for systematic writing and validation of 8,000-12,000 word paper (varies with outline detail and revision needs).


Related Skills

Prerequisite: academic-paper-strategist

  • Produces the optimized outline that this skill executes
  • Highly recommended to use first for best results

This skill can be used standalone: If you have a detailed outline from another source, you can proceed directly with this skill.

发现问题?提交给管理员复核

评分:

评论 (0)

暂无评论,成为第一个评论者吧!