Medical Research

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:

1. Main Research Themes

• What are the major topics being studied?
• What questions are researchers trying to answer?
• What approaches or methods are common?

2. Key Findings

• What are the main discoveries or results?
• What treatments or interventions show promise?
• What mechanisms are being investigated?

3. Clinical Implications

• How might this research affect patient care?
• What are practical applications?
• What changes to treatment guidelines might result?

4. Challenges and Limitations

• What problems remain unsolved?
• What resistance mechanisms exist?
• What gaps in knowledge persist?

5. 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: 1. Run: python3 .claude/skills/medical-research/pubmed_search.py "immunotherapy melanoma" 2. Analyze ~10 recent papers 3. Identify themes: resistance mechanisms, combination therapies, biomarkers 4. Write plain-language summary explaining immune checkpoint inhibitors, how they work, current challenges 5. 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: 1. Activate Skill automatically 2. Run: python3 .claude/skills/medical-research/pubmed_search.py "Alzheimer's disease pathogenesis" 3. Synthesize findings on amyloid, tau, inflammation, genetic factors 4. Explain competing theories in accessible language 5. Highlight controversies and ongoing debates 6. Save comprehensive summary

Example 3: Emerging Technology

User request: Tell me about CRISPR gene therapy research

Expected behavior: 1. Run: python3 .claude/skills/medical-research/pubmed_search.py "CRISPR gene therapy clinical trials" 2. Focus on clinical applications, safety, efficacy 3. Explain CRISPR technology simply (e.g., "molecular scissors that can edit DNA") 4. Include specific examples of diseases being treated 5. Address ethical considerations if prominent in literature 6. 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

1. Start specific: Use precise medical terms 2. Iterate if needed: Broaden or narrow based on results 3. Check paper count: Aim for 10-20 papers for comprehensive summary 4. Verify relevance: Ensure retrieved papers match the query 5. 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

Medical Research - Examples

This file contains example outputs from the Medical Research skill to demonstrate the expected format and style.

---

Example 1: Immunotherapy Research

Research Summary: Immunotherapy for Breast Cancer

Search Query: "immunotherapy breast cancer" Papers Found: 14,190 Papers Reviewed: 10 Date: 2025-12-13

Overview

Immunotherapy represents a revolutionary approach to cancer treatment that harnesses the body's own immune system to fight cancer cells. Unlike traditional chemotherapy that directly kills cancer cells, immunotherapy helps immune cells recognize and attack tumors. For breast cancer, particularly triple-negative breast cancer (TNBC), immunotherapy has emerged as a promising treatment option where other targeted therapies don't work.

TNBC is the most aggressive form of breast cancer, lacking the hormone receptors (estrogen, progesterone) and HER2 protein that other breast cancers have. This makes it harder to treat with standard hormone therapies or targeted drugs. Recent research shows that immunotherapy, especially immune checkpoint inhibitors, can help some TNBC patients, but scientists are still working to understand why some patients respond well while others don't.

This summary covers the latest research from 2025-2026, revealing both exciting advances and important challenges that researchers are working to overcome.

Key Research Themes

Theme 1: The PD-L1 Paradox

Research has uncovered a puzzling finding: patients whose tumors have high levels of a protein called PD-L1 (programmed death-ligand 1) often show poor responses to anti-PD-L1 immunotherapy drugs. This is counterintuitive because these drugs are designed to block PD-L1.

• Scientists discovered that PD-L1 can move into the cell's nucleus (the control center) where it promotes cancer spread to the lungs
• The drug interferon-gamma (IFN-γ), produced during immunotherapy, may paradoxically help cancer metastasize by activating certain pathways
• This explains why some patients with "high PD-L1" tumors don't benefit from immunotherapy and may even experience aggressive disease progression

Why it matters: Understanding this paradox could help doctors identify which patients will truly benefit from anti-PD-L1 drugs and which might need different approaches.

Theme 2: Novel Immunotherapy Approaches

Researchers are developing innovative alternatives to standard checkpoint inhibitors:

Bispecific Antibodies (BsAbs):

• These engineered antibodies can grab two different targets at once, like molecular bridges connecting immune cells to cancer cells
• Market already valued at $8.65 billion, showing strong clinical promise
• Can simultaneously block tumor growth signals and activate immune responses

Dendritic Cell Vesicles:

• Tiny bubble-like particles from immune cells that can be inhaled to prevent lung metastasis
• Target the "pre-metastatic niche" - the environment cancer cells create before spreading to new organs
• Early studies show promise in preventing rather than just treating metastatic disease

CAR-T Cells Targeting DR5:

• Genetically modified immune cells programmed to attack cancer cells through the DR5 pathway
• Bonus effect: they also suppress MDSCs (myeloid-derived suppressor cells), which normally help cancer hide from the immune system
• Showing minimal side effects in early testing

Theme 3: Understanding Resistance Mechanisms

Several biological mechanisms explain why immunotherapy sometimes fails:

NK Cell Dysfunction:

• Natural Killer (NK) cells are specialized immune cells that should kill cancer cells that hide from other immune cells
• In TNBC, a "micropeptide" (tiny protein) called UGDH-AS1 paralyzes NK cells, preventing them from attacking tumors
• The hostile tumor environment further weakens NK cell function

HIF-1 Pathway Activation:

• In low-oxygen areas of tumors, a protein called HIF-1 (hypoxia-inducible factor 1) gets activated
• HIF-1 promotes cancer stem cells (cells that can regenerate tumors) and helps cancer evade immune detection
• This pathway is linked to higher recurrence and metastasis rates

Why it matters: Identifying these resistance mechanisms provides new targets for combination therapies that could overcome treatment failure.

Theme 4: Tumor Microenvironment and Biomarkers

The environment surrounding breast tumors significantly influences immunotherapy success:

• Tumor-resident microbiome: Bacteria living in tumors (not just the gut) may affect how well immunotherapy works
• Immune-stromal heterogeneity: Analysis of over 13,000 breast tumors shows that immune cell patterns vary significantly by ancestry, affecting treatment response
• Pre-metastatic niche: Cancer cells prepare distant organs (like lungs) for future metastasis by sending signals that create welcoming environments

Why it matters: These findings suggest that personalized medicine approaches considering a patient's genetic background and tumor microbiome could improve treatment selection.

What This Means for Patients/Treatment

Current Clinical Impact:

• Immunotherapy is now standard treatment for some TNBC patients, particularly when combined with chemotherapy
• Doctors increasingly test tumors for PD-L1 levels and immune cell infiltration, though interpretation is complex
• Combination approaches (immunotherapy + chemotherapy or targeted drugs) show better results than single agents

Emerging Possibilities:

• New antibody therapies could offer options for patients who don't respond to current immunotherapies
• Inhalable treatments to prevent lung metastasis could become preventive therapy for high-risk patients
• Microbiome-modulating treatments might enhance immunotherapy effectiveness

Timeline Considerations:

• Most novel approaches (bispecific antibodies, dendritic cell vesicles, CAR-T cells) are in early clinical trials
• Expect 3-5 years before widespread availability if trials continue showing promise
• Current checkpoint inhibitors remain the primary immunotherapy option for eligible TNBC patients

What Patients Should Know:

• Not all breast cancer responds equally to immunotherapy; TNBC shows the most promise
• High PD-L1 levels don't guarantee immunotherapy success; comprehensive tumor testing is important
• Combination therapies show better results than immunotherapy alone for most patients

Challenges and Open Questions

Major Challenges:

• The PD-L1 Paradox: Why do some high-PD-L1 tumors resist anti-PD-L1 therapy? How can we identify responders beforehand?
• Immune Suppression: How can we reactivate NK cells and overcome the immunosuppressive tumor microenvironment?
• Metastasis During Treatment: Why does IFN-γ, produced during immunotherapy, sometimes promote cancer spread?
• Biomarker Complexity: Single markers like PD-L1 are insufficient; what combination of factors predicts response?

Open Research Questions:

• Can we modulate the tumor microbiome to enhance immunotherapy?
• How do ancestry and genetic background affect optimal immunotherapy approaches?
• Can we prevent the formation of pre-metastatic niches before cancer spreads?
• What combination therapies can overcome resistance mechanisms?

What Researchers Are Working On:

• Comprehensive immune profiling beyond PD-L1 testing
• Combination therapies targeting multiple resistance pathways simultaneously
• Preventive strategies for high-risk patients before metastasis occurs
• Personalized approaches based on ancestry, microbiome, and tumor immune signatures

Key Terminology

• Triple-Negative Breast Cancer (TNBC): Breast cancer lacking estrogen receptors, progesterone receptors, and HER2 protein. Most aggressive subtype, representing 15-20% of breast cancers, with highest recurrence rates.

• Immune Checkpoint Inhibitors (ICIs): Drugs that release the "brakes" on the immune system. PD-L1 and PD-1 act like brake pedals that cancer cells use to stop immune attacks; ICIs remove these brakes.

• PD-L1 (Programmed Death-Ligand 1): A protein that cancer cells display like a "don't eat me" flag to avoid immune system attacks. High levels traditionally thought to predict immunotherapy response, but research shows this is more complex.

• Tumor Microenvironment: The ecosystem surrounding a tumor, including immune cells, blood vessels, bacteria, and support cells. This environment can either help the immune system fight cancer or help cancer hide and grow.

• Bispecific Antibodies: Engineered proteins designed in labs that can simultaneously grab two different targets - typically one part grabs a cancer cell while the other grabs an immune cell, bringing them together for attack.

• Pre-Metastatic Niche: A prepared environment in distant organs (like lungs or liver) that cancer cells create before actually spreading there. Like sending advance scouts to prepare a landing site for invasion.

• CAR-T Cells: "Chimeric Antigen Receptor T cells" - a patient's own immune cells genetically modified in a lab to recognize and attack specific cancer targets, then returned to the patient's body.

Selected Recent Papers

1. Nuclear PD-L1 drives IFN-γ-promoted lung metastasis of triple-negative breast cancer via POLR2A-mediated transcriptional activation of LY6E (PMID: 41388312, 2025)

• Authors: Wang X, Zhou Q, Wang P et al.
• Key contribution: Discovered that PD-L1 in the cell nucleus (not just on the surface) promotes metastasis, explaining why high PD-L1 patients sometimes respond poorly to anti-PD-L1 therapy. Identifies LY6E as a potential therapeutic target.
• https://pubmed.ncbi.nlm.nih.gov/41388312/

2. The long non-coding RNA UGDH-AS1 encodes an NK-cell-inhibiting micropeptide in triple-negative breast cancers (PMID: 41387419, 2025)

• Authors: Zhang Z, Li F, Dai X et al.
• Key contribution: Identified a tiny protein that paralyzes natural killer cells in TNBC tumors, revealing a new mechanism of immune evasion. Suggests targeting UGDH-AS1 could restore NK cell function and enhance immunotherapy.
• https://pubmed.ncbi.nlm.nih.gov/41387419/

3. Targeting pre-metastatic niche with inhalable dendritic cell vesicles to prevent breast cancer lung metastasis (PMID: 41388289, 2025)

• Authors: Leng Q, Yu Z, Zhao Y et al.
• Key contribution: Developed an innovative inhalable treatment using immune cell vesicles to prevent lung metastasis by disrupting the pre-metastatic environment. Represents a preventive rather than reactive approach.
• https://pubmed.ncbi.nlm.nih.gov/41388289/

4. Immune-stromal heterogeneity in breast cancer across diverse ancestries: impact on prognosis and treatment response (PMID: 41387728, 2025)

• Authors: Alamukii NA, Kovács A, Raghavan S et al.
• Key contribution: Analyzed 13,000+ breast tumors showing that immune patterns vary significantly by ancestry, affecting how patients respond to immunotherapy. Highlights need for personalized approaches considering genetic background.
• https://pubmed.ncbi.nlm.nih.gov/41387728/

5. HIF-1-mediated ISG20 expression promotes breast cancer stemness and immune evasion (PMID: 41385111, 2026)

• Authors: Yang Y, Zuo Q, Ramu V et al.
• Key contribution: Revealed how low oxygen in tumors activates HIF-1, which then promotes cancer stem cells and helps tumors hide from the immune system. Identifies ISG20 as a potential therapeutic target to overcome resistance.
• https://pubmed.ncbi.nlm.nih.gov/41385111/

Summary

Immunotherapy research in breast cancer, particularly TNBC, has evolved from simple checkpoint blockade to a sophisticated understanding of complex tumor-immune interactions. The field is currently grappling with a fascinating paradox: the very marker (PD-L1) used to select patients for immunotherapy may indicate resistance in some cases due to its unexpected role in promoting metastasis.

The research landscape reveals both challenges and opportunities. On the challenge side, multiple resistance mechanisms have been identified - from micropeptides that paralyze NK cells to hypoxia-driven immune evasion pathways. These discoveries explain why single-agent immunotherapy shows limited efficacy in many patients. On the opportunity side, innovative approaches like bispecific antibodies, inhalable immune cell vesicles, and advanced CAR-T cells offer new paths forward. The recognition that tumor-resident microbiomes and ancestry-specific immune patterns affect treatment response is pushing the field toward truly personalized medicine.

The next frontier appears to be prevention rather than treatment - disrupting pre-metastatic niches before cancer spreads and using combination approaches that simultaneously address multiple resistance mechanisms. For patients, this means that while current immunotherapy options remain limited to specific TNBC subtypes, the pipeline of emerging treatments is robust. The shift from "does immunotherapy work?" to "which immunotherapy works for which patient, and how can we make it work better?" represents significant progress in this challenging disease.

---

This summary is based on scientific literature available in PubMed and is intended for educational purposes. It does not constitute medical advice. Patients should consult with their healthcare providers about treatment options appropriate for their specific situation.

---

Example 2: Disease Mechanism Research

Research Summary: Parkinson's Disease Alpha-Synuclein Aggregation

Search Query: "Parkinson's disease alpha-synuclein aggregation" Papers Found: 8,743 Papers Reviewed: 10 Date: 2025-12-13

Overview

Parkinson's disease is a progressive neurological condition affecting movement, balance, and coordination. At the heart of Parkinson's lies a mystery involving a protein called alpha-synuclein. In healthy brains, this protein helps neurons (brain cells) communicate. But in Parkinson's, alpha-synuclein misfolds - imagine a properly folded paper crane suddenly crumpling into a ball - and these misfolded proteins clump together into toxic aggregates called Lewy bodies.

Understanding why and how alpha-synuclein aggregates is crucial because these clumps appear to spread from cell to cell like a slow-moving infection, gradually destroying the dopamine-producing neurons that control movement. This research aims to prevent aggregation, stop its spread, or clear existing aggregates before they cause irreversible damage.

[Note: This is a template example - actual content would be filled in by running the PubMed search and analyzing results]

Key Research Themes

Theme 1: Mechanisms of Aggregation

[Content based on actual papers retrieved...]

Theme 2: Cell-to-Cell Spreading

[Content based on actual papers retrieved...]

Theme 3: Therapeutic Interventions

[Content based on actual papers retrieved...]

[Continue with full template structure...]

---

These examples demonstrate the expected output format, writing style, and level of detail for the Medical Research skill.

Medical Research Skill

A Claude Code skill that retrieves scientific papers from PubMed and creates accessible, plain-language research summaries.

What It Does

This skill bridges the gap between complex scientific literature and general understanding by:

1. Retrieving - Searches PubMed for relevant scientific papers on any biomedical topic 2. Analyzing - Identifies key themes, findings, and implications from multiple papers 3. Translating - Converts technical research into plain language that anyone can understand 4. Summarizing - Creates comprehensive markdown reports with clear structure and context

When to Use

The skill automatically activates when you:

• Ask about medical or scientific research (e.g., "What's the latest on diabetes treatment?")
• Request information about diseases, treatments, or clinical trials
• Want to understand scientific consensus on health topics
• Need to catch up on recent research in a specific area

Quick Start

Simply ask natural questions:

Summarize research on immunotherapy for lung cancer

What does research say about ketogenic diet for epilepsy?

Tell me about recent CRISPR gene therapy studies

Claude will automatically: 1. Use the skill to search PubMed 2. Analyze the most recent papers 3. Create a comprehensive plain-language summary 4. Save it as a markdown file in your working directory

Output Format

Each summary includes:

• Overview: What the research area is and why it matters
• Key Research Themes: Major findings organized by topic
• Clinical Implications: What this means for patients and treatment
• Challenges: Current problems and limitations
• Key Terminology: Definitions of important scientific terms
• Selected Papers: Highlighted papers with PMIDs and links
• Summary: Overall synthesis and future directions

File Naming

Summaries are saved as:

Research_Summary_{Topic}_{YYYY-MM-DD}.md

Examples:

• Research_Summary_Immunotherapy_Melanoma_2025-12-13.md
• Research_Summary_Alzheimers_Treatment_2025-12-13.md

Requirements

The skill includes a built-in pubmed_search.py script located at .claude/skills/medical-research/pubmed_search.py. This script requires:

• Python 3.x
• Biopython library (pip install biopython)

These should already be installed in the Vibe environment.

Customization

Retrieve More Papers

Edit .claude/skills/medical-research/pubmed_search.py and change the max_results parameter:

papers = search_pubmed(query, max_results=20)  # Default is 10

Set Your Email

PubMed requires an email address for API access. Update in .claude/skills/medical-research/pubmed_search.py:

Entrez.email = "your.email@example.com"

Search Strategy

For better results:

Be Specific:

• ✓ "immunotherapy for triple-negative breast cancer"
• ✗ "cancer treatment"

Use Medical Terms:

• ✓ "myocardial infarction rehabilitation"
• ✗ "heart attack recovery"

Include Context:

• ✓ "CRISPR gene editing sickle cell disease"
• ✗ "CRISPR"

Examples

Example 1: Treatment Research

Input:

Summarize research on immunotherapy for breast cancer

What Happens: 1. Searches PubMed for "immunotherapy breast cancer" 2. Finds 14,190 papers, retrieves 10 most recent 3. Identifies themes: PD-L1 paradox, resistance mechanisms, novel approaches 4. Creates 3,000+ word plain-language summary 5. Saves as Research_Summary_Immunotherapy_Breast_Cancer_2025-12-13.md

Example 2: Disease Mechanism

Input:

What does research say about Alzheimer's disease causes?

What Happens: 1. Skill activates automatically 2. Searches for "Alzheimer's disease pathogenesis" 3. Explains amyloid hypothesis, tau protein, neuroinflammation in accessible terms 4. Discusses competing theories and current debates 5. Saves comprehensive summary

Example 3: Emerging Technology

Input:

Tell me about mRNA vaccine technology

What Happens: 1. Searches "mRNA vaccine mechanism clinical trials" 2. Explains how mRNA vaccines work (using simple analogies) 3. Covers COVID-19 applications and future potential 4. Discusses safety, efficacy, and ongoing research 5. Saves detailed summary

Tips for Best Results

Getting Comprehensive Summaries

• Be specific about what aspect interests you
• Mention if you want focus on mechanisms, treatments, or clinical trials
• Specify populations if relevant (pediatric, elderly, specific conditions)

Understanding Technical Topics

• The skill will define key terms in a "Key Terminology" section
• Explanations use analogies and plain language
• Complex mechanisms are broken down step-by-step

Finding Specific Papers

• Use the "Selected Papers" section for highlighted research
• Each paper includes PMID (PubMed ID) for direct lookup
• Links provided for one-click access to PubMed

Follow-Up Questions

After receiving a summary, you can ask:

• "Can you explain the [specific mechanism] in more detail?"
• "Which of these approaches is most promising?"
• "Are there clinical trials I can look into?"

Limitations

What This Skill Does:

• Summarizes published peer-reviewed research
• Explains scientific concepts in plain language
• Identifies trends and patterns in literature
• Provides educational information

What This Skill Doesn't Do:

• Provide medical advice or treatment recommendations
• Replace consultation with healthcare providers
• Guarantee accuracy of cited research
• Cover unpublished or non-PubMed indexed research

Currency Note:

• Focuses on most recent papers (typically last 1-2 years)
• Very new research (last few weeks) may not yet be indexed
• Some emerging topics may have limited papers available

Troubleshooting

"No papers found"

Try:

• Broadening search terms (use general terms)
• Checking spelling of medical terms
• Using alternative terminology (e.g., "myocardial infarction" vs "heart attack")
• Verifying the topic has biomedical research (PubMed covers life sciences)

"Papers are too technical"

The skill should automatically translate complex findings. If unclear:

• Ask for clarification on specific terms
• Request the skill focus on clinical implications
• Check the "Key Terminology" section for definitions

"Results seem narrow"

• Modify the .claude/skills/medical-research/pubmed_search.py script to retrieve more papers
• Use broader search terms
• Ask for a follow-up search on related topics

Technical Details

How It Works

1. Query Processing: Converts your natural language question into a PubMed search query 2. Paper Retrieval: Uses Biopython's Entrez module to query PubMed API 3. Analysis: Reads titles, abstracts, authors, and metadata from retrieved papers 4. Synthesis: Identifies themes, patterns, and key findings across papers 5. Translation: Converts technical language to plain English with analogies and context 6. Structuring: Organizes information into clear sections with hierarchy 7. Documentation: Saves as markdown file with proper citations

Data Sources

• Primary: PubMed/MEDLINE (https://pubmed.ncbi.nlm.nih.gov/)
• Coverage: 35+ million biomedical literature citations
• Update Frequency: Daily; new papers added continuously
• Access: Free, public database maintained by US National Library of Medicine

Quality Assurance

Summaries are designed to:

• Write at 10th-grade reading level (accessible to general audience)
• Define all technical terms and abbreviations
• Cite specific papers with PMIDs for verification
• Note limitations and areas of uncertainty
• Include educational disclaimers
• Present balanced view including challenges

Contributing

To improve this skill:

1. Enhance the script: Modify .claude/skills/medical-research/pubmed_search.py to retrieve more metadata 2. Improve templates: Update SKILL.md with better structure suggestions 3. Add examples: Contribute sample summaries to EXAMPLES.md 4. Report issues: Note any topics that produce poor summaries

Related Resources

• PubMed: https://pubmed.ncbi.nlm.nih.gov/
• Biopython Tutorial: https://biopython.org/wiki/Documentation
• MeSH Browser: https://meshb.nlm.nih.gov/ (for finding medical search terms)
• Plain Language Guidelines: https://www.plainlanguage.gov/

Version History

• v1.0 (December 2025): Initial release
• PubMed integration via Biopython
• Plain-language summarization
• Automatic file saving
• Comprehensive templates

License

Part of the Vibe project. See main repository for license details.

---

Questions or Issues?

If you encounter problems or have suggestions for improvement, please open an issue in the Vibe repository.