When should you use it?
- You have raw genetic data and recent labs
- You want to make the most of a short appointment
- Your supplement list keeps growing
Not for: Emergencies or new severe symptoms; Changing prescriptions without your doctor
What do you need, and what do you get?
You need
- Raw genetic data export
- Recent lab report
- Imaging or scan summaries
- Supplement and medication list
You get
- One-page summary of notable values
- Variants worth asking about
- Questions for your doctor
- Supplement overlap notes
How does it work, step by step?
- Collect files locallyPut the exports in one private folder on your own machine.
- Summarize each sourceHave the agent list out-of-range lab values, scan findings and supplements with doses.
- Check genetic notes carefullyAsk it to name each variant it mentions and how strong the published evidence is, and to say when evidence is weak.
- Cross-referenceFlag where a supplement, lab value or variant might interact, as questions rather than advice.
- Write the visit briefProduce a one-page summary and a ranked list of questions to bring to your appointment.
The prompt
Our own prompt, inspired by Farzad (@farzyness) and written to work in Claude Code and Claude. Values in lime are placeholders.
Fill in the blanks (optional; the copied prompt uses your values)
I am preparing for a doctor visit. Here is what I am sharing: DNA export {dna_file}, labs {lab_report}, scans {scan_summary}, supplements {supplement_list}. 1. List lab values outside the reference range with the range shown. 2. List any genetic variants you mention by ID with a one-line note on evidence strength (strong, moderate, weak). 3. Note possible overlaps between supplements, labs and variants as questions, not advice. 4. End with the 8 most useful questions to ask my doctor. Do not suggest dose or medication changes.
Inspired by Farzad (@farzyness): My favorite use case for Claude Code (from which I stole from someone else on X):. The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.
What can go wrong?
- Consumer DNA files contain errors and many variant links are weak; treat them as leads.
- Long chats drift into confident advice; ask for sources on each claim.
Our take
Farzad (@farzyness) handed Claude Code a raw 23andMe export, comprehensive bloodwork, artery plaque scan results and full supplement routine, then talked for two hours to get a plan matched to his data. We frame the output as an organized brief and questions to take to a clinician. Having all of your own data in one conversation is the useful part. Most people walk into appointments with scattered PDFs and a supplement shelf they cannot fully explain. Lining those up into questions is a real gain, and it keeps decisions where they belong. We deliberately stop short of a personal plan, because genetic associations are often weak and a model will sound sure either way. Keep the files in a private local folder, and ask for sources and evidence strength on every claim. Never change a prescription based on the chat. Consumer DNA files contain errors, so treat any variant as a lead.
Where the time goes back: Hours of cross-checking lab ranges, variants and supplement labels become a one-page brief before an appointment.
FAQ
Can this replace my doctor?
Is it safe to upload my DNA file?
Source and credit
Inspired by Farzad (@farzyness). We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.
- Inspired by
- My favorite use case for Claude Code (from which I stole from someone else on X):
- Creator
- Farzad (@farzyness)
- Published on
- X
- Original date
- Jan 5, 2026
- License
- Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
- Originally for
- Claude Code
- AI dry run
- Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. PASS: Gates held on desk check; no live sends, bookings or purchases.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails. - Verified
- Oct 24, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.
Are you the creator? Claim or correct this page.