科研 Skill
研究计划预注册撰写
writing-research-plan
Converts an approved design into a pre-registered plan with a falsifiable hypothesis and task list. Use after brainstorming-research produces a design doc.
SKILL.md
Writing a Research Plan
Take an approved design doc (input/ideas/<slug>-design.md) and turn it into an executable research plan (input/ideas/<slug>-plan.md). The plan is the contract between idea and execution: it pre-registers the hypothesis, locks the method, lists bite-sized tasks, and defines verification criteria.
Announce at start: "Using writing-research-plan to convert the design into a pre-registered plan."
If unmet: explain to the user which condition is missing, ask for a short
reason, write it to knowledge/_meta/gate-overrides.log, and continue.
Methodology and Pre-Registration
The project frontmatter (see CLAUDE.md) declares a methodology. This
skill behaves differently depending on the value:
-
hermeneutic (default for theology / exegesis / Quellenkritik / interpretive archaeology): pre-registration is not enforced. Instead the plan requires a precise research question, a method sketch (which hermeneutic procedures, on which material), and a list of expected sources. Hypothesis revision through new reading is legitimate and gets recorded in
knowledge/_meta/log.md— the hermeneutic circle is constitutive, not a methodological failure. -
quantitative (for geostatistics, 14C Bayesian work, quantitative DH methods): full pre-registration. Hypothesis and Falsification Criteria are frozen from sign-off onward. Deviations are documented; the result is explicitly marked as exploratory rather than turned into a confirmation by silent plan edits.
-
mixed: pre-registration per sub-study. Quantitative tasks get a
pre-registered: trueblock; hermeneutic tasks don't. The plan marks this per task in the task block's frontmatter.
The plan templates below show both variants.
Checklist
- Read the design doc —
input/ideas/<slug>-design.md. If it does not exist or is not approved, stop and invokebrainstorming-research. - Lock the hypothesis — one sentence, falsifiable or argumentatively prüfbar, in the user's own words after confirmation
- Define falsification criteria — what observation would make you reject the hypothesis
- Enumerate data sources — exact sources, with status (acquire pending / on disk / already in wiki)
- Decompose into tasks — an Acquire-sources task (before ingest), then Ingest tasks, Analysis tasks, Synthesis tasks, Draft tasks; each step 2–15 min. Give every Draft task a realistic, generous word count — it is a floor for development, not a quota; too-tight targets are the main cause of dense, bullet-reflowed prose (see
drafting-manuscript) - Specify verification criteria — how we know the project is done (lint green, N sources ingested, chapter renders, …)
- Write plan doc to
input/ideas/<slug>-plan.mdusing the template below - Self-review — no placeholders, type consistency across tasks, spec coverage from design doc
- User reviews and signs off on hypothesis + task list — this is the pre-registration moment
- Hand off to executing-research-plan
Process Flow
digraph writing_plan {
"Read approved design doc" [shape=box];
"Design exists & approved?" [shape=diamond];
"Back to brainstorming" [shape=box];
"Lock hypothesis" [shape=box];
"Define falsification" [shape=box];
"Enumerate data sources" [shape=box];
"Decompose tasks" [shape=box];
"Specify verification" [shape=box];
"Write plan doc" [shape=box];
"Self-review" [shape=box];
"User pre-registers?" [shape=diamond];
"Invoke executing-research-plan" [shape=doublecircle];
"Read approved design doc" -> "Design exists & approved?";
"Design exists & approved?" -> "Back to brainstorming" [label="no"];
"Design exists & approved?" -> "Lock hypothesis" [label="yes"];
"Lock hypothesis" -> "Define falsification";
"Define falsification" -> "Enumerate data sources";
"Enumerate data sources" -> "Decompose tasks";
"Decompose tasks" -> "Specify verification";
"Specify verification" -> "Write plan doc";
"Write plan doc" -> "Self-review";
"Self-review" -> "User pre-registers?";
"User pre-registers?" -> "Write plan doc" [label="changes"];
"User pre-registers?" -> "Invoke executing-research-plan" [label="signed off"];
}
Plan Document Template
Save to input/ideas/<slug>-plan.md. The frontmatter status: value depends
on the methodology: pre-registered (quantitative or mixed with pre-registered
parts) vs. ready (hermeneutic).
Quantitative / Mixed (with pre-registration)
---
title: "<Plan title — same slug as the design>"
type: plan
created: YYYY-MM-DD
updated: YYYY-MM-DD
status: pre-registered
author: mixed
design: "[[<slug>-design]]"
---
# <Title> — Research Plan
> **For agentic workers:** REQUIRED SUB-SKILL: Use the `executing-research-plan` skill (`research-superpowers:executing-research-plan` in Claude Code) to run this plan task by task. Tasks use checkbox (`- [ ]`) syntax so each maps onto one task-list item.
**Research Question:** <precise guiding question>
**Hypothesis (pre-registered, immutable):** <one sentence>
**Falsification Criteria:** <what would falsify H>
**Stop Criterion:** <when is data collection done — sample size, time window, data availability>
**Method:** <2–3 sentences on the procedure>
**Output target:** book-chapter | article | grant-proposal | conference-talk
**Language:** de | en | mixed
---
Hermeneutic (no pre-registration)
---
title: "<Plan title — same slug as the design>"
type: plan
created: YYYY-MM-DD
updated: YYYY-MM-DD
status: ready
author: mixed
design: "[[<slug>-design]]"
---
# <Title> — Research Plan
> **For agentic workers:** REQUIRED SUB-SKILL: Use the `executing-research-plan` skill (`research-superpowers:executing-research-plan` in Claude Code) to run this plan task by task.
**Research Question:** <precise guiding question>
**Method sketch:** <which hermeneutic / source-critical / interpretive procedures, on which material>
**Expected source corpus:** <primary sources, secondary literature in keywords — may broaden through reading>
**Iteration expectation:** <which phases the plan will iterate — e.g. "repeated switching between ingest and synthesis until the source picture stabilises">
**Output target:** book-chapter | article | grant-proposal | conference-talk
**Language:** de | en | mixed
---
In both variants, Data Sources, Tasks, and Verification follow as shown
below. For mixed, quantitative tasks mark pre-registered: true in the
block frontmatter.
Data Sources
- [[finkelstein-2003]] — acquire pending (not yet on disk)
- [[mazar-2011]] — on disk, ingest pending
- Dataset:
input/data/megiddo-stratigraphy.csv - Archive: Israel Antiquities Authority, IAA-Nr. XXX (access: on-site visit required)
Tasks
Task 1: Acquire source PDFs
Files: <library>/pdf/*.pdf, input/bibliography/acquisition-todo.md
- Run
acquire-sourceson the A+B set fromliteraturguide.md(auto-downloads Open-Access PDFs, writes the manual-download worklist) - Download any paywalled originals manually (university VPN / library proxy) into
<library>/pdf/under the exact<bibkey>.pdffilename - Re-run
acquire-sourcesto reconcile until the required originals are present (or explicitly deferred)
Task 2: Ingest foundational sources
Files: knowledge/sources/*.md, output/bibtex/references.bib
- Ingest Finkelstein 2003 via
ingest-sourceskill →knowledge/sources/finkelstein-2003.md - Ingest Mazar 2011 via
ingest-sourceskill →knowledge/sources/mazar-2011.md - Verify BibTeX keys match citation style (chicago-author-date.csl)
Task 3: Analyze stratigraphic sequence
Files: output/data-analysis/stratigraphy.py, output/data-analysis/results/phases.json
- Create
stratigraphy.py— loadsinput/data/megiddo-stratigraphy.csv - Compute Bayesian phase overlaps via PyMC
- Save posterior to
results/phases.json - Write assumption-check note in
knowledge/synthesis/stratigraphy-assumptions.md
Task 4: Synthesize into a stable knowledge page
Files: knowledge/synthesis/chronologie-debatte.md
- Pull entities [[tel-megiddo]], [[tel-rehov]], [[finkelstein]], [[mazar]]
- Argument structure: Position A, Position B, own reading, evidence weight
- Cite ≥ 10 sources from BibTeX
- Set
status: stableonce reviewed by user
Task 5: Draft chapter "Forschungsstand — Chronology"
Files: output/book/text/03-forschungsstand.md
- Confirm
wiki-lintis green - Draft 2500–3500 words via
drafting-manuscriptskill - Citations
[@bibkey]fromoutput/bibtex/references.bib
Verification
The project is complete when:
-
scripts/lint-wiki.pyexits 0 - ≥ 15 sources ingested, all with
status: stableorreview -
knowledge/synthesis/chronologie-debatte.mdhasstatus: stable -
make renderinoutput/book/produces PDF without errors - Peer review round completed via
requesting-peer-review - Hypothesis explicitly confirmed, refuted, or marked "inconclusive — exploratory"
Deviations Log
Append to knowledge/_meta/log.md whenever scope, method, or data sources change after pre-registration. Reference the plan slug.
## No Placeholders (plan failures — never write these)
- "TBD", "see later", "add relevant literature"
- "more tasks as needed" (make concrete or delete)
- Tasks without file path ("an analysis somewhere")
- Citations without BibTeX key ("probably Finkelstein")
- Verification criteria without measurability ("the chapter feels good")
## Self-Review Checklist (inline)
1. **Spec coverage:** every point from `input/ideas/<slug>-design.md` — is it covered by a task?
2. **Placeholder scan:** remove the red flags (see above)
3. **Type consistency:** if task 2 writes `phases.json`, does task 3 reference exactly that file?
4. **Falsifiability:** is the hypothesis actually testable, or just a thesis?
5. **Granularity:** are all steps 2–15 min? Split longer ones
6. **No gaps between Acquire → Ingest → Analysis → Synthesis → Draft**
Fix errors inline.
## Sign-off
After self-review:
- For **quantitative** or **mixed with pre-registered tasks**:
> "Plan written and saved at `<path>`. Please explicitly confirm the
> **Hypothesis** and the **Falsification Criteria** — once confirmed they
> are frozen (pre-registered). Deviations later go into log.md. OK?"
>
> Only after explicit confirmation set `status: pre-registered` and
> invoke `executing-research-plan`.
- For **hermeneutic**:
> "Plan written and saved at `<path>`. Research question, method sketch,
> and expected sources are recorded. Hypothesis remains open — revision
> through reading is logged in `_meta/log.md`. Does the plan look right?
> Then I'll set `status: ready`."
>
> After confirmation set `status: ready` and invoke
> `executing-research-plan`.
## Key Principles
- **Hypothesis before data** — pre-registered
- **Bite-sized tasks** — 2–15 min, exact file path
- **Verification measurable** — not "good enough"
- **Deviation log** — document deviations, don't hide them
- **YAGNI** — only tasks that follow from the design
示例提示词
帮我把这份设计方案转化为带可证伪假设和任务清单的预注册研究计划。
所属插件
安装该插件会一并获得同领域的 62 个 Skill / MCP。
同插件的其他科研 Skill
- 学术文献获取下载Auto-downloads open-access PDFs for graded sources and writes a manual worklist for paywalled items. Use after literature-review, before ingest-source.
- 单篇文献入库管理Adds one verified PDF to a shared bibliography library, computing its citation key and keywords. Use to register a single document outside the full workflow.
- 数字人文计算方法Surveys computational DH methods: text mining, NLP, corpus linguistics, archive standards, GIS, network analysis, and OCR. Use for an overview of available DH tools.
- 考古地理信息系统分析Runs archaeological GIS analysis with GeoPandas: catchment analysis, viewshed, kernel density, spatial statistics. Use for mapping archaeological site locations.
- 论证结构提取可视化Extracts and visualizes argument structure: Toulmin reconstruction, support/attack graphs, circularity detection. Use to diagram an argument’s logical structure.
- 考古遗物类型学分析Performs artifact analysis: typological classification, ceramic attribute coding, lithic reduction analysis. Use when classifying excavated artifacts statistically.
- 引文与出处核查Checks citations, quotations, page numbers, and bibliography entries for accuracy and source fit. Use before submission to verify source-supported claims.
- 研究选题构思设计Explores research intent, field state, methodology, and ethics, then writes an approved design doc. Use before starting a new research question or chapter.
- 术语译名一致性检查Audits a manuscript for terms, translations, abbreviations, and names whose forms must stay stable. Use when terminology drift could undermine an argument.
- 计算语用学分析Runs computational pragmatics: speech act classification, implicature detection, discourse coherence, politeness analysis. Use for language-use-and-meaning studies.
- 文献综述学术对话构建Turns verified sources into a structured scholarly conversation among positions, not a chronological summary. Use when synthesizing literature into a debate narrative.
- 语料库语言学分析Performs corpus analysis with NLTK/spaCy: collocation, keyword analysis, KWIC concordance, word2vec, diachronic trends. Use for quantitative text-corpus research.
- 数字档案馆资源检索Queries digital archive REST APIs: Europeana, Chronicling America, Internet Archive, HathiTrust, DPLA, OAI-PMH. Use when searching or harvesting heritage metadata.
- 学术手稿章节撰写Drafts a chapter or article section by section from synthesized wiki content, agreeing each section’s argument first. Use to turn stable research synthesis into prose.
- 研究项目漂移状态检测Checks a shared source library and registered projects for drift: un-indexed PDFs, unmerged keywords, bibkey collisions. Use to audit a multi-project research setup.
- EMI课堂语料转写分析Segments and annotates classroom transcripts, coding language choice and discourse function with anonymised output. Use to quantify classroom talk or translanguaging.
- 人工智能伦理分析Analyzes AI ethics: algorithmic fairness metrics, bias auditing, value alignment, moral philosophy reasoning. Use to evaluate an AI system’s ethical dimensions.
- 研究计划任务执行Executes a ready research plan task by task, gating acquisition before ingest and dispatching analysis. Use after writing-research-plan to run the planned work.
- 研究项目结项清单Runs the final completion checklist: rendering, bibliography, wiki, data, reproducibility, archiving. Use when a manuscript is finalized and ready to publish.
- 形式逻辑自动推理Applies formal logic tools: Z3 SMT solver, modal logic S5, SAT encoding, Lean4 proof verification. Use to formalize or mechanically check a logical argument.
- 研究问题凝练构建Turns a broad topic or intuition into a precise, contestable, researchable question. Use at the start of a project when no defensible question yet exists.
- 历史地理信息系统分析Performs historical GIS work: georeferencing old maps, digitizing boundaries, temporal territory overlays, population interpolation. Use for historical spatial analysis.
- 历史语言学比较分析Analyzes historical linguistics in Python: sound correspondences, cognate detection, phylogenetic trees, Swadesh lists. Use when reconstructing language relationships.
- 历史社会网络分析Builds historical network analyses: correspondence networks, prosopography, temporal community detection, Gephi export. Use to map relationships among historical actors.
常见问题
研究计划预注册撰写能做什么?
Converts an approved design into a pre-registered plan with a falsifiable hypothesis and task list. Use after brainstorming-research produces a design doc。
如何安装研究计划预注册撰写?
在科搭 Codar 左侧点击「插件」,在「组件」页签搜索「研究计划预注册撰写」并安装;在 Claude Code 或 Codex 中可以发送本页的安装提示词,或下载 Skill 包后解压到 ~/.claude/skills 或 ~/.codex/skills。
研究计划预注册撰写需要哪些权限?
该科研 Skill没有声明额外权限;执行具体操作时仍由 Agent 按自身规则请求确认。
研究计划预注册撰写经过官方审核吗?
是。SpeedSkills 收录的每个版本都会记录内容摘要与权限声明,当前版本已通过安全审核。