agentleFS
DGU-stallionDGU-stallion/PaperPilot/skills/paper-search/SKILL.md

paper-search

Skill4 starsChanged 2 months agoMITView on GitHub

Paper search — use when the user wants to find academic literature, build a reference list, verify citations exist, or needs reference expansion from anchor papers. Looks for research direction in research_proposal.md; asks if absent. literature-review reads this output.

17 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use extract-paper-images, paper-analyze and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Clay-HHKSkill

    extract-paper-images

    从论文中提取图片,优先从arXiv源码包获取真正的论文图

    Also in 4 projects that use this

  • Clay-HHKSkill

    paper-analyze

    深度分析单篇论文,生成详细笔记和评估,图文并茂

    Also in 3 projects that use this

  • Google CloudSkill

    citation-verification

    Verify that citations in a document resolve to real records and that claimed titles match. Use before trusting a document's reference list.

    Also in 3 projects that use this

  • MicrosoftSkill

    web-design-reviewer

    This skill enables visual inspection of websites running locally or remotely to identify and fix design issues. Triggers on requests like "review website design", "check the UI", "fix the layout", "find design problems". Detects issues with responsive design, accessibility, visual consistency, and layout breakage, then performs fixes at the source code level.

    Also in 3 projects that use this

  • xfstudioSkill

    code-review-excellence

    Master effective code review practices to provide constructive feedback, catch bugs early, and foster knowledge sharing while maintaining team morale. Use when reviewing pull requests, establishing review standards, or mentoring developers.

    Also in 3 projects that use this

  • xfstudioSkill

    uv-package-manager

    Master the uv package manager for fast Python dependency management, virtual environments, and modern Python project workflows. Use when setting up Python projects, managing dependencies, or optimizing Python development workflows with uv.

    Also in 3 projects that use this

  • +2 more from Clay-HHK

The file

Files
SKILL.md

Paper Search

锚点深搜:先找高质量锚点论文并下载,从其参考文献中提取二次搜索线索,再向外扩展。

职责边界

所需研究方向 + 核心关键词
产出paper_search.md(验证后的文献清单)+ 锚点 PDF
完成锚点 4-6 篇已下载,二次扩展已执行,每条引用有验证状态
不做综述写作(→ literature-review);非学术材料(→ data-collector)

Step 1 — 收集信息

  1. 读 topics/research_proposal.md,提取:核心概念、研究问题、已提关键词
  2. 读对话中用户已提供的方向
  3. 缺失则追问(一次一个):"研究方向是什么?想搜哪些理论或关键词?"

完成:关键词(中英文)已确定。


Step 2 — 锚点深搜

2a  关键词搜索 → 选出 4-6 篇锚点论文
         ↓
2b  下载锚点 PDF(自动,不询问)
         ↓
2c  从 PDF 提取参考文献 → 二次搜索
         ↓
2d  补充搜索(填补缺口维度)

2a — 找锚点

用核心关键词搜索,按以下标准选取 4-6 篇锚点:

优先说明
期刊层级A+/A 优先
被引量>500 或领域经典
直接相关与研究问题核心关联
来源多样不能全是同一作者/团队;中英文各 ≥1 篇

完成:锚点论文 4-6 篇已选定,每篇有选取理由。

2b — 下载锚点 PDF

锚点论文自动下载,不询问用户。

下载渠道优先级(每渠道超时 30 秒):

优先级渠道工具
1Sci-Hub 镜像download_scihub(identifier=DOI, base_url=mirror)
2Semantic Scholar OAdownload_semantic(paper_id="DOI:...")
3Tavily PDF 搜索tavily_search("{title} filetype:pdf") → curl
4源原生download_arxiv / download_biorxiv 等

Sci-Hub 镜像顺序:sci-hub.st → sci-hub.ru → sci-hub.se。

save_path = "./literature/pdfs"

下载后验证:

  1. file <路径> — 必须含 PDF document
  2. 重命名为 {年份}_{简化标题}.pdf

失败处理:标记 [锚点下载失败-需手动],继续流程但在报告中提醒。

完成:锚点 PDF 已下载到 literature/pdfs/,失败的已标记。

2c — 参考文献提取与二次搜索

对每篇已下载的锚点 PDF:

  1. 提取参考文献:用 markitdown 提取 PDF 全文,定位 References/参考文献 部分
  2. 识别高相关论文:从参考文献中筛选与研究问题直接相关的论文
  3. 二次搜索:用 search_crossref 或 search_semantic 查找这些论文
  4. 元数据验证:用 get_crossref_paper_by_doi(DOI) 确认

提取命令:

markitdown "path/to/paper.pdf" | grep -A 500 -i "references\|参考文献"

或 Python:

from markitdown import MarkItDown
md = MarkItDown()
result = md.convert("path/to/paper.pdf")
# 在 result.text_content 中搜索 References 部分

提取重点:

  • 锚点论文引用的理论奠基作(如 RBV 锚点引用的 Barney, Teece 等)
  • 锚点论文引用的方法论先驱
  • 与研究问题相关的实证研究

完成:每篇锚点的参考文献已提取,高相关论文已二次搜索并验证。

2d — 补充搜索

检查当前候选池的覆盖:

  • 理论文献是否充足?
  • 方法论文献是否有?
  • 中英文是否均衡?
  • 是否有明显缺口维度?

针对缺口执行补充搜索。

完成:候选池已覆盖理论、方法、实证三类,无明显缺口。


Step 3 — 筛选

维度纳入排除
相关性与研究问题直接相关仅标题相似
时效性近 10 年(核心近 5 年)过时且无奠基价值
质量同行评审期刊/会议未审稿文章

期刊层级:A+(顶刊)、A(SSCI Q1-Q2)、B(Q2-Q3)、经典(被引>1000)。

完成:纳入列表已确定,每篇有相关度和层级标注。


Step 4 — 验证引用

禁止凭记忆生成引用。标题必须来自检索结果原文。

状态标准
✓ 已验证≥2 源一致 或 CrossRef DOI 确认
△ 单源仅 1 源但有完整 DOI
✗ 存疑信息不一致,需用户确认

完成:所有论文有验证状态,存疑已标记。


Step 5 — 写报告

paper_search.md 结构

# 文献搜索报告

## 一、搜索策略
关键词、工具、锚点深搜路径;
漏斗:检索 X 篇 → 锚点扩展 Y 篇 → 纳入 N 篇。

## 二、研究空白
现有研究的 2-3 个空白;有无竞争论文。

## 三、核心锚点论文(4-6 篇)
| # | 标题 | 作者 | 年份 | 期刊 | 层级 | 被引 | 选入理由 | PDF状态 |

## 四、二次扩展论文
从锚点参考文献中发现的重要论文。

## 五、完整文献列表
| # | 标题 | 作者 | 年份 | 期刊 | 层级 | 相关度 | 被引 | 状态 | DOI |

完成:五节均已填写,锚点有 PDF 状态标注。


Step 6 — 询问下载(非锚点)

锚点已在 Step 2b 自动下载。此步询问其他论文:

✅ 文献搜索完成

📄 literature/paper_search.md
  漏斗: 检索 X 篇 → 锚点扩展 Y 篇 → 纳入 N 篇
  锚点: [已下载 M 篇 / 失败 K 篇]

📊 研究空白: [简述]

其他 [N-M] 篇论文需要下载吗?
  A. 下载全部
  B. 指定下载
  C. 暂不下载(推荐,需要时再下)

工具分工

目标首选备选
精确查找CrossRefOpenAlex
主题探索Semantic ScholarGoogle Scholar
元数据验证CrossRef get_crossref_paper_by_doi—
中文论文Tavily site:cnki.net—
PDF 下载Sci-Hub → Semantic → Tavily—

禁止:download_with_fallback(会匹配无关文献)


质量 Checklist

  • 锚点 4-6 篇,有选取理由
  • 锚点 PDF 已下载(失败的已标记)
  • 从锚点参考文献做了二次搜索
  • 所有引用来自实际检索,标题完整
  • 每篇有验证状态和 DOI
  • 存疑引用已告知用户

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.