Skills for Real Engineers. Straight from my .agents directory.
science-study-coach-flow
Agent skill fromyzlin286-wq/Deepscient
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.
How to install science-study-coach-flow
npx skills add yzlin286-wq/Deepscient# or add only this skill to Claude Code: copy the folder
dsh-science/skills/science/science-study-coach-flow/
# into
~/.claude/skills/science-study-coach-flow/ (all projects)
.claude/skills/science-study-coach-flow/ (one project)
1# 流程指南:把"写论文"拆成十二个有门的阶段23## 为什么需要流程45研究最常见的失败不是"某一步做错了",而是**顺序错了**:67- 先有数据,再挑检验 → p 值不再有意义。8- 先招人,再算样本量 → 要么白招,要么招不满。9- 先写完讨论,再选题 → 整篇要重写。10- 先投稿,再看报告规范 → 直接退修或 desk reject。1112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13对着库里的真实状态判一遍:过、不过、还是"没留下记录"。1415## 十二个阶段1617| 阶段 | 一句话目标 | 硬门(不过就停) |18| --- | --- | --- |19| S0 选题 | 缩成一句可否证、一年做得完的话 | 课题已建、问题写成一句、学科已识别 |20| S1 文献 | 知道做到哪一步,且留检索记录 | 至少检索一次、库内文献 ≥10 |21| S2 设计 | 选一个审稿人认的设计,前置条件逐条落实 | 设计已定、假设可证伪、主结局唯一 |22| S3 样本量与预注册 | 招第一个人之前把 n 与计划锁死 | 样本量有依据 |23| S4 测量 | 每个变量落到具体测量动作 | 每个变量都有测量定义 |24| S5 采集 | 按方案采,过程留痕 | 数据进 `data/`、至少一次留痕执行 |25| S6 数据管理 | 别人能看懂、能跑起来 | 环境已锁、代码进 `code/` |26| S7 分析 | 用事前方法算,报效应量与区间 | 主分析已跑、有效应量与 CI |27| S8 图表 | 图能独立读懂、能直接排版 | 至少一张图、通过期刊自检 |28| S9 写作 | 按目标刊结构写,结论指向证据 | 稿子已建、目标刊已定、lint 无 block |29| S10 内审 | 用审稿人的方式打自己一遍 | 溯源链校验通过 |30| S11 投稿返修 | 按刊要求出件,返修逐条回应 | 四件套已出 |3132## 读它的正确方式3334`science_study_coach { projectId }` 返回三样东西,按这个顺序看:35361. **`current`** —— 你现在真正在哪个阶段。注意:它可能比你以为的早。三个门卡在 S2 却已经在写引言,37 这是最常见的状态。382. **`nextActions`** —— 只列当前阶段最该动手的几条。优先阻塞项,其次警告项,再次"没留下记录"的,39 最后才是自查项(标 `reason: self-check`)。403. **`stages`** —— 全流程一眼过。`✗` 是有阻塞,`·` 是有警告或未判定。4142## 门的状态有四种,含义不同4344| 状态 | 含义 | 该怎么办 |45| --- | --- | --- |46| `pass` | 库里查得到,且达标 | 往下走 |47| `fail` | 查到了但不达标 | 按 `fix` 修 |48| `skip` | **查不到**(还没做,或没留记录) | 去做,或补记录。**skip 不是过关** |49| `manual` | 机器判不了 | 自己去核对,核对完在心里打勾 |5051`skip` 与 `pass` 的区别是这套流程最要紧的一点。把"没查过"当"没问题",是绝大多数返修意见的来源。5253## 第一次上手的最小路径5455```56science_preflight ← 环境体检57…
Excerpt shown under the repository's MIT license.
More skills in Deepscient
- ▸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
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
An AI skill that provides design intelligence for building professional UI/UX across multiple platforms.
