#950 by GitHub stars
yzlin286-wq/Deepscient
让 AI 与人类一起攻克真正的科学难题 —— 12 阶段 55 门 · 36 工具 · 98 条技能 · 证据链与可复现
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
dsh-science/skills/science/science-agent-scheduling/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-campaign-design/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-consensus-report/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-discipline-methods/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-experiment-runbook/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-figure-for-journal/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-frontier-to-paper/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-fulltext-hunting/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-journal-intel/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-journal-selection/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-literature-search/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-manuscript-formatting/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-org-roles/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-rebuttal-playbook/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-reporting-standards/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-reproducibility/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-source-matrix/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-statistics-choice/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-stats-report/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-study-coach-flow/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
dsh-science/skills/science/science-top-journal-writing/SKILL.md
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 088b0228df4f
No earlier version was observed for comparison.
Heat
8.2- Star growth, last 7 daysAhead of 100% of tracked repositories
- Star growth, last 24 hoursAhead of 100% of tracked repositories
- Rank movement, last 7 daysNot available
- Commits, last 30 daysAhead of 86% of tracked repositories
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekAhead of 0% of tracked repositories
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
- science-agent-scheduling
dsh-science/skills/science/science-agent-scheduling - science-campaign-design
dsh-science/skills/science/science-campaign-design - science-consensus-report
dsh-science/skills/science/science-consensus-report - science-discipline-methods
dsh-science/skills/science/science-discipline-methods - science-experiment-runbook
dsh-science/skills/science/science-experiment-runbook - science-figure-for-journal
dsh-science/skills/science/science-figure-for-journal - science-frontier-to-paper
dsh-science/skills/science/science-frontier-to-paper - science-fulltext-hunting
dsh-science/skills/science/science-fulltext-hunting - science-journal-intel
dsh-science/skills/science/science-journal-intel - science-journal-selection
dsh-science/skills/science/science-journal-selection - science-literature-search
dsh-science/skills/science/science-literature-search - science-manuscript-formatting
dsh-science/skills/science/science-manuscript-formatting - science-org-roles
dsh-science/skills/science/science-org-roles - science-rebuttal-playbook
dsh-science/skills/science/science-rebuttal-playbook - science-reporting-standards
dsh-science/skills/science/science-reporting-standards - science-reproducibility
dsh-science/skills/science/science-reproducibility - science-source-matrix
dsh-science/skills/science/science-source-matrix - science-statistics-choice
dsh-science/skills/science/science-statistics-choice - science-stats-report
dsh-science/skills/science/science-stats-report - science-study-coach-flow
dsh-science/skills/science/science-study-coach-flow - science-top-journal-writing
dsh-science/skills/science/science-top-journal-writing
How it works
science-agent-scheduling
When your agent uses it
Use when a campaign is running and work must be dispatched, retried or throttled — when a shard sits in running with no heartbeat, when a lease expired and the result was duplicated, when retries storm the source APIs, when budget or concurrency must be cut mid-campaign, or when someone wants to fix a stuck task by editing state directly.
What it loads into context
Frontmatter
79 tokensAlways available to the agent
SKILL.md
1,421 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 一条总原则
- 租约(lease)
- 心跳
- 幂等键
- 退避
- 背压
- 预算熔断
- 失败时先看事件流
- 禁止清单
- 收尾检查
Files
SKILL.mdcore4,865 bytesSKILL.mdcore4,865 bytes
Read SKILL.md
# 任务调度:租约、幂等与背压
## 一条总原则
**调度器是唯一能改状态的人。** 模型、脚本、人都不直接写 task 状态;
所有状态变化只由 `science_campaign_run` 的动作(claim / heartbeat / complete / fail /
release)产生。手工改 state 会让事件流与状态表对不上,之后每一次重跑都不可信。
## 租约(lease)
- 领取任务即获得租约,`leaseMs` 默认 120000(2 分钟)。
- 租约不是所有权,是**临时独占**;到期未续,任务自动回到 pending,任何人可再领。
- 领取前必须写 `owner`、`leaseUntil`、`attempt`。三项缺一,这次领取无效。
- 续租用 `heartbeat`,间隔取 `leaseMs/3`(约 40 秒)。不要等到快到期才续。View the complete source on GitHubscience-campaign-design
When your agent uses it
Use when a research goal is too large for one agent turn and must be split into schedulable work — when planning a campaign across S0–S11, when deciding how fine a shard should be (literature, statistics, figures, writing), when setting gates and vote counts so parallel results can be trusted, or when a batch plan produced duplicates, idle workers or unusable outputs.
What it loads into context
Frontmatter
88 tokensAlways available to the agent
SKILL.md
1,667 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 四层拆解
- 拆到哪一层为止
- 分片粒度经验值
- Shard 的最小字段
- Gate 怎么设
- votes 怎么定
- 预算与失败
- 反模式
- 交付前自检
Files
SKILL.mdcore5,461 bytes
Read SKILL.md
# Campaign 设计:从研究问题到可分片的任务
## 四层拆解
```
Campaign(一个课题的一次冲顶)
└─ Workstream(S0–S11 中的一段,有独立 gate)
└─ Task(一个可验收的交付,有 owner 与 gate)
└─ Shard(一次模型调用能做完的最小单元)
```
- **Campaign** 与课题一一对应,有唯一 `campaignId`;跨课题不复用。
- **Workstream** 必须有 `stage`(S0…S11)与 `gate`;没有 gate 的 Workstream 不允许存在。
- **Task** 是派单单位,必须有唯一 owner(L2 岗位)与 `acceptance`。View the complete source on GitHubscience-consensus-report
When your agent uses it
Use when several agents or votes produced different answers and the result must be merged into one defensible statement — when running the three-vote rule, when weighting preprints against peer-reviewed evidence, when escalating an unresolved disagreement, or when writing the evidence table and gap list that a campaign report must carry.
What it loads into context
Frontmatter
69 tokensAlways available to the agent
SKILL.md
1,574 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 三票规则
- 证据权重
- 冲突怎么升级
- 报告结构
- 缺口怎么写
- 三条纪律
Files
SKILL.mdcore5,228 bytes
Read SKILL.md
# 共识与报告:把多份结果合成一个可辩护的结论
## 三票规则
- 默认 `votes = 3`,`voteThreshold = 0.66`:三票中至少两票一致才采用。
- 三票必须**相互独立**:不同分片、不同输入顺序、不得共享中间摘要。
同一上下文里连问三次不算三票。
- **只对结论型判断投票**:纳入/排除、新颖性、审稿意见分级、字段抽取。
统计量、图表、格式转换属于确定性产物,一票即可,多票是浪费。
- 三票全不同 → 不合并、不取平均,直接按"分歧"上报。
- 两票一致但第三票给出反例证据 → 不按多数,转人工/上级裁决。
## 证据权重
View the complete source on GitHubscience-discipline-methods
When your agent uses it
Use when a research question is still a question rather than a plan — when you know the field but not which study design the field accepts, how many samples it demands, which measurements count as evidence, or which software is standard; also use when a review came back saying the design cannot answer the question, when you are switching fields or advising a student in an unfamiliar one, when two supervisors disagree about method, or when you need to know what reviewers in that field attack first before you commit months to a protocol.
What it loads into context
Frontmatter
113 tokensAlways available to the agent
SKILL.md
1,761 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 这一层解决什么问题
- 入口
- 一个学科包里有九样东西
- 怎么用(四步)
- 与其他层的分工
- 常见误用
Files
SKILL.mdcore6,514 bytes
Read SKILL.md
# 学科方法库:先定设计,再谈统计
## 这一层解决什么问题
一个研究生最常见的卡点不是「不会做统计」,而是**问错了问题类型**。
他手里有一个现象、一份数据、一个想说的话,但不知道这个领域**用什么方式把现象变成证据**。
心理学靠随机分配与操纵检查,经济学靠外生变异与平行趋势,化学靠表征链,材料靠批次与标准,
机器学习靠多种子与强基线——同样是「证明 A 影响 B」,各领域要求的证据形态完全不同。
所以这一层的产出不是结论,是**这个领域公认的走路方式**:
能问什么、用哪个设计、要多少样本、量什么、跑什么方法、装什么软件、审稿人先打哪里。
## 入口View the complete source on GitHubscience-experiment-runbook
When your agent uses it
Use when turning an approved plan into an actual run — when the design and sample size are settled and someone needs to produce data, execute code, or reproduce a result; when a run must be recordable (inputs, outputs, environment, seed) so that six months later the numbers in the paper can be traced back; when a result cannot be reproduced and the difference has to be explained rather than waved away; also use when deciding where raw data, analysis code and generated artifacts belong.
What it loads into context
Frontmatter
108 tokensAlways available to the agent
SKILL.md
1,102 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 这一层要守住的三件事
- 顺序
- `science_exp_run` 的字段怎么填
- 失败了怎么办
- 复现性检查(交给别人之前)
- 常见错误
Files
SKILL.mdcore4,003 bytes
Read SKILL.md
# 从方案到一次可复现的执行
## 这一层要守住的三件事
1. **原始数据只进 `data/`,只读。**
清洗结果放 `data/clean/`,分析产物放 `experiments/<slug>/<runId>/`。
一旦在原始文件上直接改,你永远回不到采集当天的状态——而审稿人问的第一个问题往往就是那个状态。
2. **每次执行都要能被别人重跑。**
`science_exp_run` 记的是:脚本内容指纹、输入文件 sha256、输出文件 sha256、
Python 与关键库版本、seed、是否允许网络。缺任何一项,这次执行就只是"你跑过一次"。
3. **Python 层默认无网。**
数据下载在采集阶段做完,分析阶段不该联网。要联网就显式 `network: true`,
记录里会写明——这是给自己留的证据,不是给别人看的仪式。
View the complete source on GitHubscience-figure-for-journal
When your agent uses it
Use when a figure has to survive peer review and print — when choosing a figure type for a comparison, a distribution, a paired design or a survival curve; when a journal demands specific column widths, minimum type size, vector formats or resolutions; when error bars must state what they are; when a figure came back from review as "unreadable" or "insufficient resolution"; also use when deciding whether a bar chart is the wrong choice.
What it loads into context
Frontmatter
101 tokensAlways available to the agent
SKILL.md
1,404 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 尺寸先定,字号其次,字体最后
- 图型选择:别用柱状图默认一切
- 误差棒:必须声明是什么
- 颜色
- 文件格式
- 自检(`ready` 字段)
- 常见退修理由与对策
Files
SKILL.mdcore4,501 bytes
Read SKILL.md
# 图:先决定它要证明什么,再决定它长什么样
## 尺寸先定,字号其次,字体最后
期刊的栏宽是**物理量**:单栏 85–90 mm,双栏 180–190 mm。你的图必须按这个物理尺寸排版,
然后再去看字号。
反过来做(先画一张 8 英寸宽的图,再在排版时缩到 85 mm)会出人命:缩放 2.4 倍后,
8 pt 的字变成 3.3 pt,印刷出来是一团灰。
`science_figure` 就是按这个顺序做的:`widthMm` → `fontSize`(默认 7 pt,低于 6 pt 直接拒绝)→ 出件。
| 预设 | 毫米 | 什么时候用 |
| --- | --- | --- |View the complete source on GitHubscience-frontier-to-paper
When your agent uses it
Use when you want to turn a fast-moving implementation from GitHub — a new optimizer, a training trick, a scaling recipe — into a publishable contribution, and you need to find which of them is actually worth reading, extract what it really claims, test whether the field has already filled that slot, and decide what evidence a band-3/4/5 venue would demand; also use when a repo reports a number but no mechanism and you need to know what is missing before writing.
What it loads into context
Frontmatter
109 tokensAlways available to the agent
SKILL.md
4,064 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 为什么这一层存在
- 一、scan:先解决「读什么」
- 二、distill:把散文拆成可证伪的句子
- 三、gap:缺口不是「没人做」,是「这五类里缺哪类」
- 四、design:band 3 和 band 5 要的东西不是一回事
- 五、plan:结构,不是措辞
- 六、常见误读
- 七、一条完整链
Files
SKILL.mdcore14,148 bytes
Read SKILL.md
# 从 GitHub 前沿到顶刊:把实现变成可证伪的主张
## 为什么这一层存在
一个新优化器的 README 通常写着「1.35x speedup」。这不是 claim,是**观察**。
顶刊要的是「通过什么机制导致什么,以及在什么范围内成立」。
两者之间的距离,就是这一层要量出来的东西。
五个工具是一条链,每一步的产物是下一步的输入:
```
science_frontier_scan GitHub 上有什么值得读 → frontier_item
science_knowledge_distill README 里到底声称了什么 → knowledge_claim
science_gap_matrix 这个方向被填了多少 → gap_entryView the complete source on GitHubscience-fulltext-hunting
When your agent uses it
Use when you have a DOI/PMID/arXiv id and need the actual PDF or JATS full text, when a download returns HTML instead of a PDF, or when building the methods section and you need page-level quotes; covers the nine-stage fallback chain, how to verify a real PDF, and what provenance to record.
What it loads into context
Frontmatter
77 tokensAlways available to the agent
SKILL.md
1,002 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 铁律:200 不等于拿到了全文
- 九级顺序(不跳级,除非上层明确 no-oa)
- 拿到之后
- 失败时怎么汇报
Files
SKILL.mdcore3,258 bytes
Read SKILL.md
# 全文获取:九级回退与验真
## 铁律:200 不等于拿到了全文
实测中大量站点回 **200 + `text/html`**:Cookie 门、登录页、"请启用 cookie"跳转页。
**每一级都必须验**:`content-type` 含 `pdf` **且** 文件头是 `%PDF-`;JATS 则验根元素 `<article`。
判失败的特征:
- 内容是 HTML 且含 `cookies_not_supported` / `sign in` / `access denied` / `captcha`
- 重定向后的 `finalUrl` 落在 `/login`、`/sorry/`、`?error=`
- 长度 < 20KB 的"PDF"(多为占位页)
## 九级顺序(不跳级,除非上层明确 no-oa)
View the complete source on GitHubscience-journal-intel
When your agent uses it
Use when you need to know how a journal actually behaves rather than what its aims-and-scope page claims — its real subfield mix over the last two years, first-round review duration and comment difficulty from authors, Chinese-author review speed and accept-rate wording; also use when a manuscript was rejected for "scope" and you need evidence about which journals genuinely publish that kind of work.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
1,841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 为什么需要这一层
- 一、用 topicFootprint 判断「这是不是我的主场」
- 二、用 reviewReality 排时间表,不用它做取舍
- 三、用 chinaIntel 看中国作者的真实反馈
- 四、和 science_journal_fit 配合
- 五、全量补数据
- 六、常见误读
Files
SKILL.mdcore6,324 bytes
Read SKILL.md
# 期刊情报:看它发过什么,不看它说自己喜欢什么
## 为什么需要这一层
`aims & scope` 是营销文案。真正决定送审的是**编辑桌上那堆稿子的分布**。
一本刊过去两年在你这个子领域发了 0 篇,你的 cover letter 写得再好也过不了初筛;
发了 9%,那就是它的主场之一。这个差别 scope 关键词永远告诉不了你。
`science_journal_intel` 把三个独立来源拼成一张卡:
| 源 | 回答的问题 | 记的是什么 |
|---|---|---|
| OpenAlex | 它**实际**在发什么 | 近两年每篇论文的 primary topic,按 subfield 汇总占比 |
| SciRev | 作者**实际**经历了什么 | 首轮月数、总量月数、即时拒天数、审稿轮数、意见难度、样本数 |View the complete source on GitHubscience-journal-selection
When your agent uses it
Use when choosing where to submit, when a paper was desk-rejected without review, when deciding between a top journal and a solid sub-journal, or when planning the submission chain; covers the three-tier strategy, why scope beats prestige, format red lines, and how to use transfer and portable-review routes.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,227 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 第一原则:desk reject 的头号原因不是「不够好」,是「投错刊」
- 第二原则:格式红线比内容更早杀死你
- 第三原则:三个加速通道比"写得更好"更省时间
- 第四原则:投稿顺序是一次决策,不是一连串反应
- 硬规矩
Files
SKILL.mdcore4,447 bytes
Read SKILL.md
# 选刊:先算 desk reject 概率,再谈影响因子
## 第一原则:desk reject 的头号原因不是「不够好」,是「投错刊」
编辑初筛以分钟计,只回答一个问题:**这篇稿子在这个刊的读者群里成立吗?**
所以选刊的第一步永远是 scope 匹配,不是影响因子排序。
用 `science_journal_fit` 跑一次,看三档:
- **reach**:档位更高 **且有主题命中**。没有主题命中的综合刊不会进 reach——那不是冲刺,是碰运气。
- **target**:档位匹配、分数够。正常情况下这里是你的第一投。
- **safe**:降 1–3 档且设计兼容。用于时间压力(毕业、结题、抢 priority 窗口)。
`science_journal_fit` 的分数可以逐项拆开,投之前先读它给你的 `matchedTopics` / `missingTopics`:View the complete source on GitHubscience-literature-search
When your agent uses it
Use when starting a literature review, checking whether an idea is already published, or updating a topic search; covers query construction, multi-source coverage, deduplication, inclusion criteria, and what must be recorded so the search is reproducible.
What it loads into context
Frontmatter
57 tokensAlways available to the agent
SKILL.md
908 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 何时用
- 前置检查
- 步骤
- 产出与落库
- 常见失败与降级
Files
SKILL.mdcore3,005 bytesSKILL.mdcore3,005 bytesSKILL.mdcore3,005 bytes
Read SKILL.md
# 文献检索(可复现的检索策略)
## 何时用
- 课题开局要做现状梳理;
- 有 idea 需要做新颖性(scoop)检查;
- 距离上次检索超过 3 个月,需要增量更新。
## 前置检查
1. `science_preflight` 已跑过,数据根与技能就位;
2. 确认检索源:`mcp__paper__*`、`mcp__arxiv__*`、`mcp__zotero__*` 是否可用;不可用则退化为 `web_search` + `web_fetch` 打公开 API(Crossref / OpenAlex / Semantic Scholar / PubMed E-utilities),并在结果里标注"覆盖度下降"。
3. 与用户确认三件事:**时间窗**、**语言**、**纳排标准**(哪些不算:预印本?综述?非同行评议?)。
## 步骤
1. **拆概念块**:把课题拆成 2–4 个概念块(PICO 或"方法 + 对象 + 指标"),每个块列同义词、缩写、上下位词。View the complete source on GitHubscience-manuscript-formatting
When your agent uses it
Use when a manuscript draft needs to become a submittable file — you have the content (from science_paper_plan or your own draft) and now need it in the target journal's actual layout; also use when a submission was desk-rejected for formatting (word count over, too many figures, uncited references, wrong reference style), when the same paper must be retargeted from one journal to another without rewriting, or when a submission system demands DOCX or JATS rather than PDF.
What it loads into context
Frontmatter
112 tokensAlways available to the agent
SKILL.md
2,791 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 为什么这一层是机械的(也正因如此才能自动化)
- 一、先建稿,别从空文件开始
- 二、写完先查要件,别先出 PDF
- 三、出件
- 四、看环境
- 五、支持哪些刊
- 六、和 M4 的接法
Files
SKILL.mdcore9,305 bytes
Read SKILL.md
# 排版与出件:一份源,四个产物,换刊只改一行
## 为什么这一层是机械的(也正因如此才能自动化)
写作里有一部分判断只有作者能做:这个 claim 成不成立、这个 gap 真的存在吗、这条审稿意见怎么回。
这些在 M2/M4 里,靠 `science_paper_audit` 和 `science_gap_matrix`。
剩下那一部分是**纯机械的**:字数有没有超、图有几张、引用的 key 在文献表里有没有、
参考文献该用 `[1]` 还是 `(Jordan et al., 2024)`、页边距多少毫米、要不要连续行号。
这些东西完全由**目标期刊**决定,跟内容无关——所以只要你把「刊」告诉机器,它就该自己算完。
这就是 `paper.md` 的设计前提:**正文里不写任何格式**。
```markdownView the complete source on GitHubscience-org-roles
When your agent uses it
Use when deciding who should do a piece of research work, when a task has no owner or two owners, when a role's authority is unclear (who may approve a design, who may veto an analysis), or when a disagreement must be escalated; covers the 13-role RACI table across S0–S11, the six authority types, the veto bypass to the PI, and the escalation path from shard worker to PI.
What it loads into context
Frontmatter
97 tokensAlways available to the agent
SKILL.md
1,639 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 何时用
- 四层结构
- RACI:13 岗位 × 8 个阶段
- 六种权限的边界
- 升级与 veto
- 何时派谁
- 三条硬规则
Files
SKILL.mdcore5,402 bytes
Read SKILL.md
# 科研机构岗位与权限
## 何时用
- 一件事不知道该派谁:检索、设计、跑数、作图、写稿、返修各有人。
- 两方结论冲突,需要知道谁能拍板、谁能否决。
- 一个任务被"大家都可以做"拖住,需要指定唯一负责人。
- 要在 `subagent` 里点名 `agent_preset`,需要确认该岗位对应哪个 preset。
## 四层结构
| 层 | 名称 | 一句话 | 容量 |
| --- | --- | --- | --- |
| L0 | 课题主持人 | 定方向、拍终稿、对外负责 | 1 |View the complete source on GitHubscience-rebuttal-playbook
When your agent uses it
Use immediately after receiving reviewer reports, when a revision decision says major revision, or when you disagree with a reviewer; covers the numbered-response structure, how to split fixable from arguable points, why unprincipled compliance backfires, and how to escalate a genuine disagreement to the editor.
What it loads into context
Frontmatter
69 tokensAlways available to the agent
SKILL.md
1,056 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 第一步:切条与分类(不要直接开始写)
- 第二步:逐条回应的四段结构
- 第三步:三个反直觉的规则
- 第四步:分歧升级到编辑的写法
- 硬规矩
Files
SKILL.mdcore3,804 bytes
Read SKILL.md
# 审稿回应:先分类,再逐条
## 第一步:切条与分类(不要直接开始写)
把评审意见粘进 `science_rebuttal`,它会切成编号条目并按类型标注:
| 类型 | 特征用语 | 做法 |
|---|---|---|
| **new-experiment** | additional experiments / should test / in vivo | 能补就补;不能补降级为 limitation 并说明不影响主结论 |
| **statistics** | sample size / power / CI / multiple comparison | 一律重跑,给效应量 + 95% CI + 校正方法 |
| **claim-strength** | overclaim / not supported / too strong | 降低表述强度,并在讨论里给替代解释 |
| **prior-work** | should cite / already reported | 补引用并写清增量;确属已知就改成"确认并扩展" |
| **clarity** | unclear / confusing / rewrite | 重写该段,回应里引改后原句 + page/line |
| **method-detail** | provide more detail / reagent / version | 补到可复现粒度,移到 SI 并在正文加指引 |View the complete source on GitHubscience-reporting-standards
When your agent uses it
Use when preparing a submission, when a reviewer asks for CONSORT/PRISMA/STROBE/ARRIVE/TRIPOD items, when the journal demands data/code availability statements, or when an audit reports blocking findings; covers which standard applies to which design and what each one actually requires.
What it loads into context
Frontmatter
70 tokensAlways available to the agent
SKILL.md
1,133 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 先选对规范(选错等于白写)
- 任何投稿都要有的四段声明
- 统计报告:审稿人最先查的五件事
- TOP Guidelines 8 项(透明度分级,多数刊已采纳)
- 用工具的顺序
- 硬规矩
Files
SKILL.mdcore3,865 bytes
Read SKILL.md
# 报告规范:审稿人手里的核对表
## 先选对规范(选错等于白写)
| 你的研究 | 规范 | 关键条目(缺了必被要求补) |
|---|---|---|
| 随机对照试验 | **CONSORT** | 注册号、随机化方法与分配隐藏、盲法、主要结局、受试者流程、样本量估算、ITT、不良事件 |
| 系统综述 / meta | **PRISMA** | 方案预注册(PROSPERO)、数据库与检索式、筛选流程、偏倚风险、异质性、发表偏倚 |
| 队列 / 病例对照 / 横断面 | **STROBE** | 设计类型、纳排标准、混杂处理、缺失数据、敏感性分析 |
| 动物实验 | **ARRIVE 2.0** | 物种品系、饲养条件、人道终点、随机分组(含笼效应)、评估设盲 |
| 预测模型 | **TRIPOD** | 预测因子定义、区分度**与**校准、内部/外部验证、EPV 或最小样本量 |
| 质性研究 | COREQ / SRQR | 研究者立场、抽样策略、资料饱和、编码与可信度 |
## 任何投稿都要有的四段声明View the complete source on GitHubscience-reproducibility
When your agent uses it
Use before reporting any experiment result, when a result cannot be re-run, when writing the methods section, or when preparing a reproducibility package; covers run provenance, environment capture, seed handling, artifact hashing, and the clean-room re-run check.
What it loads into context
Frontmatter
60 tokensAlways available to the agent
SKILL.md
745 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 何时用
- 前置检查
- 步骤
- 产出与落库
- 常见失败与降级
Files
SKILL.mdcore2,736 bytes
Read SKILL.md
# 可复现性(运行留痕与干净重跑)
## 何时用
- 每次要对外报告一个实验结论之前;
- 方法学部分动笔之前(先把该记的记下来);
- 有人质疑结果、或自己发现两次运行不一致时。
## 前置检查
1. 项目是否已 `science_project_open`(没有 project_id 的运行记录没有归属);
2. 实验计划是否已 `science_exp_plan`(预注册式:假设、变量、n、分析计划、风险);
3. 工作目录是否在 git 仓库内(能拿到 commit 才能谈复现)。
## 步骤
1. **固定环境**:记录解释器版本、关键依赖版本(`pip freeze` / `renv.lock` / `package-lock.json`)并算一个依赖哈希;不要只写"用了 pandas"。View the complete source on GitHubscience-source-matrix
When your agent uses it
Use when choosing which retrieval source to query, when a search returns suspiciously few results, when OpenAlex budget runs low, or when a source is rate-limited or blocked; covers per-source capability, quotas, fallback order, and how to tell "source broke" from "query too narrow".
What it loads into context
Frontmatter
69 tokensAlways available to the agent
SKILL.md
1,056 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 开局先做两件事
- 源的能力分工
- 换源顺序(预算或限流时)
- 结果太少时,先分清是谁的问题
- 硬规矩
Files
SKILL.mdcore3,452 bytesSKILL.mdcore3,452 bytesSKILL.mdcore3,452 bytesSKILL.mdcore3,452 bytesSKILL.mdcore3,452 bytes
Read SKILL.md
# 检索源矩阵:谁查什么、什么时候换源
## 开局先做两件事
1. `science_source_status` —— 看谁在冷却、OpenAlex 还剩多少美元预算;
2. `science_lit_query` —— 本地库有没有现成的。**同一课题不要重复烧配额**。
## 源的能力分工
| 源 | 强在哪 | 弱在哪 | 用法 |
|---|---|---|---|
| **OpenAlex** | 覆盖面最广,带引用数、OA 位置、摘要倒排索引 | **有美元预算**(无 key $0.10/天) | 首选,但一次查询别超 25 条 |
| **Crossref** | DOI 权威、元数据最规范、出版信息全 | 无摘要(部分有)、检索是相关性排序非布尔 | 补 DOI 与卷期页;查不到某篇的准确出版信息时用它 |
| **Europe PMC** | 生物医学全覆盖,**直接给 OA 全文链接** | 非生物医学覆盖弱 | 医学/生物方向第一顺位;M1.1 全文链的第 4–5 级靠它 |
| **PubMed** | 临床与生物医学索引最严(MeSH) | 两步请求、无 OA 链接 | 系统综述的纳排标准靠它;配合 Europe PMC 拿全文 |View the complete source on GitHubscience-statistics-choice
When your agent uses it
Use when you have data and must decide which test to run — when someone asks "该用什么检验", when a reviewer says the analysis does not match the design, when assumptions are violated and you need the honest fallback instead of the convenient one, when the effect size or correction method is missing from the results, or when a p-value is being asked to carry a claim it cannot carry; also use before data collection to size the study, and when interpreting a result that is significant but tiny.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
1,993 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 三条不能破的铁律
- 决策树怎么走
- 选检验时必须一起决定的四件事
- 假设检查:怎么查、查不过怎么办
- 本应用能直接跑的
- 结果怎么写
- 常见误用清单
Files
SKILL.mdcore6,541 bytes
Read SKILL.md
# 统计选择:设计决定检验,假设决定能不能用
## 三条不能破的铁律
1. **检验由设计决定,不由数据长什么样决定。**
先有 `design`(几组、配对否、结局类型、是否重复测量),才有候选检验。
拿到数据后试三种检验挑 p 最小的,那不是分析,是搜索。
2. **假设不满足时必须降级,不能默认继续。**
每个检验都带 `assumptions` 与 `fallback`。假设没查或不满足,正确动作是
**换 fallback 检验并在正文说明**,而不是「反正结果差不多」。
3. **p 值不是效应量。** 报告必须同时给:检验统计量、自由度、p、效应量、效应量的置信区间。
缺效应量的结果在顶刊会被直接退回补充。
## 决策树怎么走View the complete source on GitHubscience-stats-report
When your agent uses it
Use when a statistical result has to leave the analysis tool and enter a manuscript — when writing the statistics paragraph, the table notes, or the figure caption that carries a p-value; when someone asks how to report a test, an effect size, or a confidence interval; when a result is significant but the assumption check failed; when a chi-square had Yates correction or not and the value differs between tools; also use when a reviewer asks for the numbers behind a claim.
What it loads into context
Frontmatter
104 tokensAlways available to the agent
SKILL.md
1,373 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 一句话标准
- 模板(照抄结构,别照抄数字)
- 不同检验的报法
- 卡方:Yates 校正是必须先声明的
- 假设不满足时怎么报
- 多重比较
- 不该出现的句子
- 交付前自查
Files
SKILL.mdcore4,337 bytes
Read SKILL.md
# 把统计结果写成审稿人认的句子
## 一句话标准
一个结果要能独立成立,必须同时给出:**样本量 → 统计量 + 自由度 → p → 效应量 + 置信区间 → 用的是哪个检验、为什么**。
缺任何一项,这一段就会被要求补充;缺效应量最常见。
## 模板(照抄结构,别照抄数字)
```
Welch 双样本 t 检验比较咖啡因组(n = 24,M = 421 ms,SD = 52)与安慰剂组
(n = 24,M = 452 ms,SD = 55)的简单反应时,两组差异未达显著,
t(44.79) = −1.95,p = .058,Cohen's d = −0.57(95% CI −1.15 至 0.02)。
```View the complete source on GitHubscience-study-coach-flow
When your agent uses it
Use when someone is starting research and does not know the order of operations — a first-year student asking "从哪开始", "这一步要做什么", "我这样算不算做完了"; when a project feels stuck and it is unclear which prerequisite is actually missing; when explaining what a complete study requires end to end; when deciding whether the next move is reading, planning, collecting, analysing or writing. Also use before writing a Methods section, to prove every step actually happened.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
1,451 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 为什么需要流程
- 十二个阶段
- 读它的正确方式
- 门的状态有四种,含义不同
- 第一次上手的最小路径
- 卡住的时候怎么问
- 三条全程规则
Files
SKILL.mdcore5,104 bytes
Read SKILL.md
# 流程指南:把"写论文"拆成十二个有门的阶段
## 为什么需要流程
研究最常见的失败不是"某一步做错了",而是**顺序错了**:
- 先有数据,再挑检验 → p 值不再有意义。
- 先招人,再算样本量 → 要么白招,要么招不满。
- 先写完讨论,再选题 → 整篇要重写。
- 先投稿,再看报告规范 → 直接退修或 desk reject。
流程的价值不是仪式,而是让「还没做」在系统里有名字。`science_study_coach` 会把每个阶段的门
对着库里的真实状态判一遍:过、不过、还是"没留下记录"。
View the complete source on GitHubscience-top-journal-writing
When your agent uses it
Use when drafting or restructuring a manuscript for a high-impact journal, when a paper was rejected for "limited conceptual advance" or "not of broad interest", or when the abstract and title need to carry the claim; covers the C-C-C structure, the five-sentence abstract, figure-first ordering, and the Introduction funnel.
What it loads into context
Frontmatter
75 tokensAlways available to the agent
SKILL.md
1,191 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 唯一的判据:标题是不是一个断言
- C-C-C 递归:每个尺度都做同一件事
- 摘要 5 句(写完就有 80% 的送审率差异)
- Figure-first:先定 Figure 1 的一句话
- Introduction 漏斗(三段,不超过 4 段)
- Discussion:给边界,不复述
- 送审前最后一次通读的四个检查
- 硬规矩
Files
SKILL.mdcore4,229 bytes
Read SKILL.md
# 顶刊写作:让编辑在 60 秒内看懂你在主张什么
## 唯一的判据:标题是不是一个断言
编辑初筛只做一件事——读标题和摘要,判断"这篇稿子提出了什么主张,这个主张重要吗"。
- ✗ `A Study of Chaperone Effects on Protein Folding`(话题)
- ✓ `Chaperone concentration sets the folding rate of globular proteins`(断言)
断言的三个条件:有主语、有方向、可被证伪。满足这三条的标题,摘要和正文会自动变得好写,因为你知道了要证明什么。
## C-C-C 递归:每个尺度都做同一件事
**context → content → conclusion**,在四个尺度上重复:View the complete source on GitHubSkills included
- science-agent-scheduling
- science-campaign-design
- science-consensus-report
- science-discipline-methods
- science-experiment-runbook
- science-figure-for-journal
- science-frontier-to-paper
- science-fulltext-hunting
- science-journal-intel
- science-journal-selection
- science-literature-search
- science-manuscript-formatting
- science-org-roles
- science-rebuttal-playbook
- science-reporting-standards
- science-reproducibility
- science-source-matrix
- science-statistics-choice
- science-stats-report
- science-study-coach-flow
- science-top-journal-writing
Repository facts
- License
- MIT
- Created
- Oct 21, 2026
- Last pushed
- Oct 25, 2026
- Forks
- 12
- Open issues
- 0
- Project activity
- Active
Rank over the last 12 weeks
Now #950 · best #950
- #999 on Nov 5
- #957 on Nov 6
- #950 since Nov 7
Before you install
Instructions onlyAutomated signals, not a guarantee
Checked Nov 5, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 7 notes
dsh-science/skills/science/science-agent-scheduling/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-literature-search/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-literature-search/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-source-matrix/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-source-matrix/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-source-matrix/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzeddsh-science/skills/science/science-source-matrix/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- No dangerous shell commands
- 3 network referencesin 1 file: science-fulltext-hunting/SKILL.md:25
- No hidden or obfuscated text
- No instruction overrides
- No executable scripts
- No access to secrets
- No requests to vote in Hot
Show all 3 signals by file
dsh-science/skills/science/science-fulltext-hunting/SKILL.md
- Network accessline 25
6. `PMC efetch db=pmc&id={pmcid}&retmode=xml` —— 无需 key 7. 直链模板(这四个实测无门): - arXiv `https://arxiv.org/pdf/{id}`(支持 Range) - Network accessline 26
7. 直链模板(这四个实测无门): - arXiv `https://arxiv.org/pdf/{id}`(支持 Range) - PLOS `journals.plos.org/plosone/article/file?id={doi}&type=printable` - Network accessline 27
- arXiv `https://arxiv.org/pdf/{id}`(支持 Range)
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Signal confidence: 3 facts
- Network access in prose of the SKILL.mdscience-fulltext-hunting/SKILL.md:25Low confidence
- Network access in prose of the SKILL.mdscience-fulltext-hunting/SKILL.md:26Low confidence
- Network access in prose of the SKILL.mdscience-fulltext-hunting/SKILL.md:27Low confidence
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md dsh-science/skills/science/science-agent-scheduling/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
Add this badge to your README
It updates every night and links to this page.
[](https://hotskills.dev/skill/yzlin286-wq-deepscient/)