survey
Skill85 starsChanged 5 months agoMITView on GitHub
从 wiki 知识生成论文 Related Work 章节:主题分组 → 叙事结构 → LaTeX 输出,遵循 citation-verification 和 academic-writing
Files
/survey
基于 wiki 已有知识,生成可直接用于论文的 Related Work 章节。 从 wiki/papers/、concepts/、topics/ 取材,按研究方向分组(非逐篇罗列), 每组以「与本文的区别」收尾。引用遵循 citation-verification.md, 写作遵循 academic-writing.md 的 Related Work 规则。 支持 LaTeX 和 Markdown 两种输出格式。
Inputs
query:以下之一:- 研究问题描述(自由文本,如 "parameter-efficient fine-tuning for LLMs")
- claim slugs 列表(从 wiki/claims/ 中,用于围绕特定 claims 组织相关工作)
- PAPER_PLAN.md 路径(从中提取 Related Work section 定义)
--format(可选,默认latex):输出格式latex:\cite{key}引用,可直接嵌入论文markdown:[[slug]]wikilink 引用,用于 wiki 存档
--max-papers(可选,默认 30):最多引用的论文数量
Outputs
wiki/outputs/related-work-{slug}-{date}.md— Related Work 文本(归档)wiki/graph/edges.jsonl— derived_from 边(若新建 output)wiki/log.md— 追加日志- 终端输出 — Related Work 正文(方便直接复制)
Wiki Interaction
Reads
wiki/papers/*.md— Problem、Key idea、Results、Related、My takewiki/concepts/*.md— Definition、Variants、Comparison、Known limitationswiki/topics/*.md— Overview、Timeline、Open problems、Seminal workswiki/claims/*.md— Statement、source_papers(若输入为 claim slugs)wiki/ideas/*.md— Motivation(了解本文定位)wiki/index.md— 内容目录,按 importance 筛选wiki/graph/context_brief.md— 全局上下文wiki/graph/edges.jsonl— 论文间语义关系(same_problem_as、similar_method_to、complementary_to、builds_on、compares_against、improves_on、challenges、surveys).claude/skills/shared-references/academic-writing.md— Related Work 写作规则.claude/skills/shared-references/citation-verification.md— 引用纪律
Writes
wiki/outputs/related-work-{slug}-{date}.md— 归档文件wiki/graph/edges.jsonl— derived_from 边wiki/log.md— 追加操作日志
Graph edges created
derived_from:related-work output → source papers
Workflow
前置:确认工作目录为 wiki 项目根(包含 wiki/、raw/、tools/ 的目录)。
Step 1: 定位相关知识
- 解析输入:
- 若为自由文本:提取关键词,在 wiki/index.md 中匹配 tags 和 titles
- 若为 claim slugs:读取每个 claim 的 source_papers,收集相关论文
- 若为 PAPER_PLAN 路径:读取 Related Work section 的 groupings 和 citations
- 读取 wiki/graph/context_brief.md 获取全局上下文
- 读取 wiki/graph/edges.jsonl:提取论文间语义关系(same_problem_as、similar_method_to、complementary_to、builds_on、compares_against、improves_on、challenges、surveys)
- 生成候选论文列表:
- 从 index.md 按 importance 降序排列
- 按 tags 和 domain 匹配度排序
- 限制为
--max-papers篇
- 若候选论文 < 5 篇:警告「相关论文不足,建议先 /ingest 更多论文」
Step 2: 精读相关页面
对候选论文列表中的每篇论文:
- 读取
wiki/papers/{slug}.md:重点读 Problem、Key idea、Results、My take - 读取该论文关联的
wiki/concepts/*.md:重点读 Definition、Variants、Comparison - 读取相关
wiki/topics/*.md:重点读 Timeline、Open problems
记录每篇论文的:
- 核心贡献(一句话)
- 方法类别(属于哪个研究方向)
- 与本文的关系(same problem / similar method / complementary / builds on / compares against / improves on / challenges / surveys)
- 局限性(从 Limitations 或 My take 提取)
Step 3: 主题分组
遵循 shared-references/academic-writing.md 的 Related Work 规则:
- 按研究方向分组(非逐篇罗列):
- 从 wiki/topics/ 和 concepts/ 的分类中提取自然分组
- 每组 3-8 篇论文
- 分组标题描述研究方向(如 "Parameter-Efficient Fine-Tuning"),非单篇论文
- 确定组间顺序:
- 从宏观到具体(大方向 → 子方向 → 最相关方法)
- 或按时间线(早期奠基 → 发展 → 最新)
- 确定组内顺序:
- 按年份升序(展现演进)
- 重要论文详写(2-3 句),次要论文简写(1 句)
- 标注每组与本文的关系:
- 每组结尾一句:「Unlike these approaches, our method...」或「We build upon X by...」
Step 4: 生成段落
遵循 shared-references/academic-writing.md:
- 每组一段或两段:
- 开头:该方向的背景和重要性
- 中间:按组内顺序展开每篇论文的贡献
- 结尾:与本文的定位关系(必须有)
- 引用格式:
--format latex:\cite{key},key 从 citation-verification.md 的命名规则生成--format markdown:[[slug]]
- 写作规范:
- 不写平铺列表("X did Y. Z did W.")
- 每段有 topic sentence
- 使用对比连接词("While X focuses on..., Y addresses...")
- 不使用 AI 特征词汇(参考 academic-writing.md de-AI 列表)
- De-AI polish:
- 扫描并替换 AI 特征词汇
- 变化 sentence openings
- 移除 filler sentences
Step 5: BibTeX 准备(仅 --format latex)
若输出格式为 LaTeX,遵循 shared-references/citation-verification.md:
- 收集所有
\cite{key}引用 - 对每个 key,尝试获取 BibTeX:DBLP → CrossRef → S2
- 已验证的:记录 BibTeX
- 未验证的:标记
[UNCONFIRMED] - 输出 BibTeX 条目列表(可追加到 paper/references.bib)
- 报告 citation coverage
Step 6: 归档
- 生成 slug:
python3 tools/research_wiki.py slug "<query-keywords>" - 写入归档文件:
创建
wiki/outputs/related-work-{slug}-{date}.md:
正文为完整 Related Work 文本。 若 latex 格式:附录包含 BibTeX 条目。--- title: "Related Work: {topic}" type: related-work format: {latex|markdown} paper_count: {N} date_generated: YYYY-MM-DD --- - 添加 graph edges:
# output → 每篇引用的论文 python3 tools/research_wiki.py add-edge wiki/ \ --from "outputs/related-work-{slug}-{date}" --to "papers/{paper-slug}" \ --type derived_from --evidence "Cited in related work section" - 追加日志:
python3 tools/research_wiki.py log wiki/ \ "survey | {topic} | {N} papers, {G} groups, format: {format}" - 终端输出:完整 Related Work 正文 + citation coverage 统计
Constraints
- 只引用 wiki 中已有论文:不凭空编造引用。每个
\cite{}或[[slug]]必须对应 wiki/papers/ 中的页面 - 按主题分组,非逐篇列表:每段覆盖一个研究方向,非「Paper A did X. Paper B did Y.」
- 每组必须有定位句:与本文的关系(结尾处说明区别或继承)
- 候选论文 < 5 篇时警告:提示用户先 /ingest 更多论文
- BibTeX 遵循 citation-verification.md:不从 LLM 记忆生成(仅 --format latex)
- de-AI polish 必选:生成后必须执行 polish pass
- 归档到 outputs/:不直接修改 wiki 的 papers/concepts/topics 页面
- graph edges 使用 tools/research_wiki.py:不手动编辑 edges.jsonl
Error Handling
- wiki 论文不足 3 篇:报错,建议先 /ingest 足够数量的论文
- 无匹配论文:扩大搜索范围(放宽 tag 匹配),若仍无则报错
- BibTeX 获取全部失败(latex 格式):使用 [UNCONFIRMED] 占位,报告数量
- PAPER_PLAN 格式不匹配:忽略 plan 的分组建议,使用自动分组
- slug 冲突:追加日期后缀
Dependencies
Tools(via Bash)
python3 tools/research_wiki.py slug "<title>"— 生成 slugpython3 tools/research_wiki.py add-edge wiki/ ...— 添加 graph edgepython3 tools/research_wiki.py log wiki/ "<message>"— 追加日志python3 tools/fetch_s2.py search "<title>"— BibTeX fallback(S2 搜索)
MCP Servers
- 无(survey 不需要 Review LLM,可通过 /review --focus writing 单独审查)
Claude Code Native
Read— 读取 wiki 页面Glob— 查找 wiki 页面WebFetch— DBLP / CrossRef BibTeX 获取(仅 --format latex)
Shared References
.claude/skills/shared-references/academic-writing.md— Related Work 写作规则 + de-AI polish.claude/skills/shared-references/citation-verification.md— BibTeX 获取和 [UNCONFIRMED] 协议
Called by
/paper-draftStep 3(Related Work section 可委托此 skill)- 用户手动调用
Other versions21 projects use a version of this.
Word for word · 0
None found.
Edited · 4
- DAIR-AutoResearch/AutoSci★ 1.7k12% same
- PKU-DAIR/AutoSci★ 1.7k12% same
- skyllwt/AutoSci★ 1.7k12% same
- skyllwt/OmegaWiki★ 1.7k12% same
Wrote their own version · 17
- genkovich/sdd★ 187same name
- SNL-UCSB/literature-survey-skill★ 93same name
- spacedock-dev/spacedock★ 81same name
- V-Songbird/foreman★ 35same name
- jmylchreest/aide★ 19same name
- jamestexas/agents★ 2same name
- LazyerIJ/self-clone★ 2same name
- lucasacoutinho/dds★ 2same name
- svyatov/handrail★ 2same name
- gihwan-dev/claude-code-gui★ 1same name
7 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
Used alongsideProjects with it also use exp-pilot-eval, exp-pilot-run and 2 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.
DAIR-AutoResearchSkillexp-pilot-eval
Pilot result evaluation — read pilot results, apply success criteria, update idea page (pilot_result, failure_reason if failed), generate PILOT_VERDICT_REPORT. Called by /ideate Phase 5 after /exp-pilot-run.
Also in 4 projects that use this
DAIR-AutoResearchSkillexp-pilot-run
Pilot experiment execution — read Pilot Spec YAML, write pilot code, run experiment(Confirm with the user before operation and require the applicant to conduct manual inspection), return results. Called by /ideate Phase 5. Does NOT modify wiki pages or judge pass/fail.
Also in 4 projects that use this
phanghonghaoSkillposter
HTML/CSS A4 海报生成。从沙龙/活动 Markdown 文件生成科技风 A4 海报 PDF/PNG。Use when user mentions: poster, 海报, 海报生成, make poster, generate poster, 活动海报, 沙龙海报.
Also in 4 projects that use this
- +1 more from DAIR-AutoResearch
In this repository30 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- askSkill
- checkSkill
- daily-arxivSkill
- discoverSkill
- editSkill
- empirical-designSkill
- empirical-ingestSkill
- exp-designSkill
- exp-evalSkill
- exp-runSkill
- exp-statusSkill
- ideateSkill
- ingestSkill
- initSkill
- noveltySkill
- paper-compileSkill
- paper-draftSkill
- paper-planSkill
- prefillSkill
- rebuttalSkill
- refineSkill
- researchSkill
- resetSkill
- reviewSkill
- setupSkill
- stata-planSkill
- theory-ingestSkill
- variable-mapSkill
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
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.

