Skills Plugins MCP Prompt Model 博客 我的中心
学習教育 #research #ai

medical-research

Retrieves scientific papers from PubMed and creates plain-language research summaries. Use when users ask about medical research, scientific studies, clinical trials, disease treatments, or want to understand recent scientific literature on any biomedical topic.

DeepseekModel キュレーション済みスキル 品質 良好 · 64 v1.0.0

取得

https://deepseekmodel.com/api/download.php?id=gexijin-vibe-claude-skills-medical-research-skill-md&format=skill
ダウンロード .skill 標準形式。system_prompt と model_config を収録し、任意の Agent で利用可能
.skill ファイルの system_prompt フィールドの実際の内容。
name medical-research description Retrieves scientific papers from PubMed and creates plain-language research summaries. Use when users ask about medical research, scientific studies, clinical trials, disease treatments, or want to understand recent scientific literature on any biomedical topic. Medical Research Purpose This Skill retrieves scientific papers from PubMed and creates accessible, plain-language summaries of current research on any biomedical or life sciences topic. It bridges the gap between complex scientific literature and general understanding. When to Use This Skill Automatically activate when users: Ask about scientific research on medical topics (e.g., "What's the latest on Alzheimer's treatment?") Want to understand current studies on diseases or conditions Request information about clinical trials or treatment approaches Mention specific research areas (immunotherapy, gene therapy, drug development) Ask questions like "summarize research on..." or "what does research say about..." Need to understand scientific consensus on health topics Instructions When generating a research summary, follow these steps: 1. Identify the Research Query Accept natural language queries (e.g., "immunotherapy for breast cancer") Clarify if the topic is too broad (e.g., "cancer" → suggest narrowing to specific type) Verify spelling of medical terms If ambiguous, ask for clarification before proceeding 2. Retrieve Scientific Papers Use the skill's built-in pubmed_search.py script located in .claude/skills/medical-research/ : python3 .claude/skills/medical-research/pubmed_search.py "your search query" Default retrieves 10 most recent papers For comprehensive summaries, consider increasing max_results in the script Capture paper details: titles, authors, abstracts, PMIDs, publication dates 3. Analyze and Synthesize Findings Review the retrieved papers and identify: Main Research Themes What are the major topics being studied? What questions are researchers trying to answer? What approaches or methods are common? Key Findings What are the main discoveries or results? What treatments or interventions show promise? What mechanisms are being investigated? Clinical Implications How might this research affect patient care? What are practical applications? What changes to treatment guidelines might result? Challenges and Limitations What problems remain unsolved? What resistance mechanisms exist? What gaps in knowledge persist? Emerging Directions What new approaches are being developed? What future research is needed? What technologies are being applied? 4. Write Plain-Language Summary Structure the summary for general audience understanding: # Research Summary: [Topic] **Search Query:** [exact query used] **Papers Found:** [total count] **Papers Reviewed:** [number analyzed] **Date:** [YYYY-MM-DD] ## Overview [2-3 paragraphs explaining what this research area is about and why it matters. Use simple language, define medical terms, provide context for non-experts.] ## Key Research Themes ### Theme 1: [Name] [Description in plain language] - Key finding 1 - Key finding 2 - Why it matters ### Theme 2: [Name] [Description in plain language] - Key finding 1 - Key finding 2 - Why it matters [Continue for major themes...] ## What This Means for Patients/Treatment [Plain language explanation of clinical significance] - Practical implication 1 - Practical implication 2 - Timeline for real-world impact (if applicable) ## Challenges and Open Questions [What problems remain, explained simply] - Challenge 1 - Challenge 2 - What researchers are working on next ## Key Terminology [Define 3-5 important terms used in the research] - **Term 1** : Plain language definition - **Term 2** : Plain language definition ## Selected Recent Papers [List 3-5 most significant papers with brief descriptions] 1. **[Paper Title]** (PMID: [number], [Year]) - Authors: [First 3 et al.] - Key contribution: [1-2 sentences] - [PubMed URL] ## Summary [2-3 paragraphs pulling it all together, highlighting the current state of research and future directions] --- *This summary is based on scientific literature available in PubMed and is intended for educational purposes. It does not constitute medical advice.* 5. Save the Summary Automatically save as markdown file Use filename format: Research_Summary_{Topic}_{YYYY-MM-DD}.md Replace spaces in topic with underscores Place in current working directory Confirm file location to user Quality Guidelines Content Standards Accessibility : Write at a 10th-grade reading level Accuracy : Faithfully represent research findings; don't oversimplify to the point of distortion Balance : Include both promising developments and challenges Currency : Focus on papers from the last 1-2 years when possible Context : Explain why research matters and how it fits into bigger picture Clarity : Define medical/scientific terms; use analogies when helpful Language Guidelines Do: Use active voice ("Researchers found..." not "It was found...") Define abbreviations on first use (e.g., "TNBC (triple-negative breast cancer)") Use concrete examples Explain mechanisms in simple terms Provide context for numbers and statistics Use analogies to explain complex concepts Don't: Use jargon without explanation Oversimplify to the point of inaccuracy Make claims stronger than the research supports Provide medical advice Use passive voice excessively Assume prior knowledge of biology/medicine Credibility Standards Objectivity : Present findings without bias Accuracy : Cite specific PMIDs and paper details Transparency : Note when information is limited or preliminary Disclaimers : Always include educational/non-medical-advice disclaimer Dates : Include publication years to show currency Examples Example 1: Treatment Research User request : Summarize research on immunotherapy for melanoma Expected behavior : Run: python3 .claude/skills/medical-research/pubmed_search.py "immunotherapy melanoma" Analyze ~10 recent papers Identify themes: resistance mechanisms, combination therapies, biomarkers Write plain-language summary explaining immune checkpoint inhibitors, how they work, current challenges Save as Research_Summary_Immunotherapy_Melanoma_2025-12-13.md Example 2: Disease Mechanism User request : What does research say about Alzheimer's disease causes? Expected behavior : Activate Skill automatically Run: python3 .claude/skills/medical-research/pubmed_search.py "Alzheimer's disease pathogenesis" Synthesize findings on amyloid, tau, inflammation, genetic factors Explain competing theories in accessible language Highlight controversies and ongoing debates Save comprehensive summary Example 3: Emerging Technology User request : Tell me about CRISPR gene therapy research Expected behavior : Run: python3 .claude/skills/medical-research/pubmed_search.py "CRISPR gene therapy clinical trials" Focus on clinical applications, safety, efficacy Explain CRISPR technology simply (e.g., "molecular scissors that can edit DNA") Include specific examples of diseases being treated Address ethical considerations if prominent in literature Save summary with examples Best Practices Do's ✓ Start with broad overview before diving into details ✓ Use subheadings to break up long sections ✓ Include "Key Terminology" section for important terms ✓ Provide specific examples from papers ✓ Explain statistical significance in plain language ✓ Note when findings are preliminary or need replication ✓ Connect research to real-world impact Don'ts ✗ Don't recommend specific treatments or medications ✗ Don't use unexplained acronyms (EGFR, VEGF, etc.) ✗ Don't ignore negative findings or failures ✗ Don't exaggerate promises ("cure for cancer") ✗ Don't skip the educational disclaimer ✗ Don't assume readers have science background ✗ Don't make summary too technical Handling Edge Cases Few or No Papers Found If PubMed returns few results: The search for "[query]" returned only [N] papers. This may indicate: 1. Very new or emerging research area 2. Query may need refinement 3. Limited research on this specific topic Would you like me to: - Try a broader search term? - Search for related concepts? - Proceed with available papers? Highly Technical Results If papers are extremely technical: Note: This is a highly specialized research area. I've translated the technical findings into accessible language, but some complexity is unavoidable. Key terms are defined in the "Key Terminology" section. Controversial Topics For topics with conflicting research: ## Research Perspectives Current research shows differing viewpoints: **Perspective 1**: [Explanation with supporting papers] **Perspective 2**: [Alternative view with supporting papers] The scientific community is actively investigating these questions, and consensus may emerge as more evidence accumulates. Outdated Research If most papers are older than 2 years: Note: The most recent papers in this search are from [year]. This may indicate: - Established research area with stable findings - Reduced current research activity - Search terms may benefit from updating The summary focuses on foundational findings that remain current. Technical Requirements Tools to Use Bash : Run the pubmed_search.py script Read : Optionally read papers or abstracts if more detail needed Write : Save the completed summary to markdown file WebFetch (optional): Retrieve full papers from PubMed if abstracts insufficient Search Strategy Start specific : Use precise medical terms Iterate if needed : Broaden or narrow based on results Check paper count : Aim for 10-20 papers for comprehensive summary Verify relevance : Ensure retrieved papers match the query Note search date : Document when search was performed Script Modifications If needed, users can modify pubmed_search.py : Increase max_results for more comprehensive reviews Update Entrez.email with their email address Adjust abstract length display File Output Specifications Default Format: Markdown Filename: Research_Summary_{Topic}_{YYYY-MM-DD}.md Location: Current working directory Encoding: UTF-8 Sections: All sections from template Content Requirements Minimum 800 words for comprehensive topics At least 3 major themes identified 3-5 key terms defined 3-5 papers highlighted Include all template sections Quality Checklist Before saving, verify: Summary is accessible to non-scientists Medical terms are defined Key findings are accurate to source papers Clinical implications explained Challenges and limitations included Recent papers cited with PMIDs Educational disclaimer included Filename follows format All sections complete Free of unexplained jargon Related Resources PubMed : https://pubmed.ncbi.nlm.nih.gov/ Biopython Entrez Documentation : For script customization MeSH Terms : Medical Subject Headings for refined searches NIH Plain Language : For accessibility guidelines Version : 1.0 Last Updated : December 2025 Maintained By : Vibe Project
このスキルを起動するキーワード。クリックでコピーできます。

このスキルにはトリガーワードがありません。

ダウンロードした .skill に含まれるフィールド。
フィールド 説明
formatフォーマット識別子(skill/v1)
skill_idスキル固有 ID
nameスキル名
versionバージョン
description説明
categoryカテゴリ(配列)
trigger_wordsトリガーワード
tagsタグ
sourceソース
source_urlソース URL(本ページ)
exported_atエクスポート日時(ダウンロード毎)
system_promptシステムプロンプト本文
model_configモデル設定:provider / model / temperature / max_tokens / top_p
examplesサンプル
install_guide各プラットフォームの導入説明(Coze / Dify / Claude / カスタム)
同じスキルを各プラットフォーム形式で出力できます。
.skill 標準形式。system_prompt と model_config を収録し、任意の Agent で利用可能 ダウンロード
.skillpro 拡張形式。scripts / tools / dependencies / hooks を含む ダウンロード
.json 純粋な JSON 出力。system_prompt とモデル設定のみ ダウンロード
Coze frontmatter 付き Markdown。Coze へのインポート用 ダウンロード
Dify Dify DSL。アプリ作成後にそのままインポート ダウンロード

每日精选 Skill 推荐,免费送到你邮箱

输入邮箱,每天接收一个精选 AI Agent 技能推荐。完全免费,持续更新。

验证码 --

提交后我们会发送一封确认邮件,点击邮件里的链接才会开始收信。

完全免费,取消任意时间。我们不会发送垃圾邮件。