domain-modeling
Build and sharpen a medical research project's domain model. Use when the user wants to pin down research terminology or a ubiquitous language, record a consequential methodological decision, or when another skill needs to maintain the domain model.
- License 许可
- MIT
- 最近更新
- 2026-11-01
- Stars
- ★ 2
- Forks
- 0
01 前言 Front Matter
- name
- domain-modeling
- description
- Build and sharpen a medical research project's domain model. Use when the user wants to pin down research terminology or a ubiquitous language, record a consequential methodological decision, or when another skill needs to maintain the domain model.
02 资源文件 3
- 参考文档agents/openai.yaml260 B
- 参考文档CONTEXT-FORMAT.md2.7 KB
- 参考文档DECISION-FORMAT.md1.7 KB
03 SKILL.md 正文
Domain Modeling
Actively build and sharpen the research project's domain model as you design. This is the active discipline — challenging terms, inventing edge-case scenarios, and writing the glossary and decisions down the moment they crystallise. (Merely reading CONTEXT.md for vocabulary is not this skill — that's a one-line habit any skill can do. This skill is for when you're changing the model, not just consuming it.)
File structure
Most research projects have a single context:
/
├── CONTEXT.md
├── docs/
│ └── decisions/
│ ├── 0001-define-index-time.md
│ └── 0002-select-primary-estimand.md
└── analysis/
If a CONTEXT-MAP.md exists at the root, the project has multiple contexts. The map points to where each one lives:
/
├── CONTEXT-MAP.md
├── docs/
│ └── decisions/ ← project-wide decisions
├── studies/
│ └── cohort/
│ ├── CONTEXT.md
│ └── docs/decisions/ ← study-specific decisions
└── reviews/
└── meta-analysis/
├── CONTEXT.md
└── docs/decisions/
Create files lazily — only when you have something to write. If no CONTEXT.md exists, create one when the first term is resolved. If no docs/decisions/ exists, create it when the first research decision record is needed.
During the session
Challenge against the glossary
When the user uses a term that conflicts with the existing language in CONTEXT.md, call it out immediately. "Your glossary defines 'baseline' as the covariate-assessment period, but you seem to mean the index date — which is it?"
Sharpen fuzzy language
When the user uses vague or overloaded terms, propose a precise canonical term. "You're saying 'cohort' — do you mean the target population, study population, or analysis population? Those are different things."
Discuss concrete scenarios
When domain relationships are being discussed, stress-test them with specific scenarios. Invent scenarios that probe edge cases and force the user to be precise about the boundaries between concepts. For reviews and meta-analyses, include multiple reports from one cohort, multi-arm trials, repeated time points, composite outcomes, and several eligible effect estimates.
Cross-reference with research artifacts
When the user states how something works, check whether the protocol, registration, data dictionary, extraction form, statistical analysis plan, analysis code, tables, and manuscript agree. If you find a contradiction, surface it: "Your protocol starts follow-up at treatment initiation, but the analysis uses enrollment — which is right?"
Update CONTEXT.md inline
Before the first edit in a session, show the proposed canonical changes and ask for authorization to update the research files. Once authorized, when a term is resolved, update CONTEXT.md right there. Don't batch these up — capture them as they happen. Use the format in CONTEXT-FORMAT.md.
CONTEXT.md should be totally devoid of operational details. Do not treat CONTEXT.md as a protocol, data dictionary, statistical analysis plan, or a repository for procedural decisions. It is a glossary and nothing else.
Offer research decision records sparingly
Only offer to create a research decision record when all three are true:
- Hard to reverse — the cost of changing your mind later is meaningful
- Surprising without context — a future collaborator will wonder "why did they do it this way?"
- The result of a real trade-off — there were genuine alternatives and you picked one for specific reasons
If any of the three is missing, skip the decision record. Use the format in DECISION-FORMAT.md.
04 同仓库更多 Skills
grill-meta-analysis-topic
Validate a proposed meta-analysis topic against the published evidence.
teach
Teach the user a medical research skill or concept, within this workspace.
grilling
Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.
grill-me
A relentless interview to sharpen a plan or design.
screen-pubmed-abstracts
Screen a PubMed plain-text export for a systematic review or meta-analysis and produce a balanced CSV shortlist.
research
Investigate a medical research question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched, biomedical or methodological facts gathered, or reading legwork delegated to a background agent.