Opalite medical translator
Translate clinical text between clinician jargon and patient-friendly language (or across spoken languages), preserving every dosage and numeric value verbatim, expanding all abbreviations correctly, and flagging anything that needs human verification.From its SKILL.md
npx -y skills add riteshkew/yc-skills --skill opalite-medical-translatorAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
5.0 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it
Workflow
When this skill triggers, follow these steps in order.
Step 1 — Ingest Clinical Text and Target Audience
Ask the user for (or read from the prompt):
- Clinical text — the source document: discharge summary, after-visit note, lab result, prescription instructions, or any clinical snippet.
- Target audience — one of:
patient-friendly(plain English, reading level ~6th grade)caregiver(slightly more technical, but still jargon-free)spanish-patient(patient-friendly, translated into Spanish)clinician(reverse direction — patient description back into clinical shorthand)
- Language override (optional) — any ISO language name, e.g. "Mandarin", "Portuguese". Overrides the audience language if set.
If the clinical text or target audience is missing, prompt:
"Please provide: (1) the clinical text you want translated, and (2) the target audience (patient-friendly, caregiver, spanish-patient, or clinician)."
Step 2 — Extract and Inventory Critical Values
Before writing a single word of translation, internally build an inventory:
- Dosages and concentrations: e.g.
500mg,0.9% NaCl,10 units - Frequencies: e.g.
TID,BID,q8h,PRN - Durations and intervals: e.g.
x7d,x14 days,2 wks - Thresholds and reference values: e.g.
>38.5°C,<90 mmHg,HbA1c >7% - Numeric lab values: e.g.
WBC 12.4,Cr 1.1 mg/dL,INR 2.3 - Procedure and anatomy abbreviations: e.g.
s/p lap chole,RUQ,PCP,D/C
Record every item. These are the verbatim preservation list — none may be paraphrased or altered in the output.
Step 3 — Translate to the Target Register
Produce the translated document following these rules:
For patient-friendly / caregiver / language targets:
- Expand every abbreviation on first use, e.g.
TID (three times a day). - Write dosages and frequencies as:
amoxicillin 500 mg (500mg), taken three times a day (TID)— preserve the original value in parentheses if space allows, or inline directly. - Never substitute a numeric value:
500mgstays500mg, never "half a gram".38.5°Cstays38.5°C, not "slightly above normal". - Use short sentences (≤20 words each).
- Avoid Latin phrases; use plain equivalents (
by mouthnotper os). - Organise the output with clear section headers: What happened, Your medications, Follow-up appointments, Warning signs — return to the ER if you notice any of these.
- Keep the Warning signs section visually prominent (bold header).
For clinician (reverse) direction:
- Convert plain-language descriptions back to standard clinical shorthand.
- Preserve all numeric values exactly as the patient stated them.
- Flag any values that appear clinically implausible with
[VERIFY].
Step 4 — Flag Ambiguities and Add the Verification Notice
After the translated document, always append two sections:
Things to verify with your provider
List every term or instruction from the source that is:
- Ambiguous or context-dependent (e.g. a PRN medication with no stated threshold)
- An abbreviation that could have multiple expansions (e.g.
MS= multiple sclerosis OR morphine sulfate) - A value outside typical patient reference ranges that warrants explanation
- Any instruction requiring professional clinical judgement
Format as a numbered list. If nothing is ambiguous, write: "No ambiguous terms identified — still confirm all instructions with your provider before acting."
Verification notice
Always end with this exact block, adapted for the target language if translating:
Important: This translation is a communication aid only. It is not medical advice. Always confirm your medications, doses, and follow-up instructions directly with your healthcare provider or pharmacist before making any changes to your care.
Step 5 — Quality Check
Before returning output, verify the inventory from Step 2:
- Every dosage value from the source appears verbatim in the output (500mg not "500 milligrams" unless the original said that).
- Every frequency abbreviation is expanded AND the original abbreviation is preserved in parentheses on first use.
- Every duration and threshold appears unchanged.
- The Things to verify with your provider list is present.
- The Important verification notice is present.
- The
## Rubricsection is appended.
If any check fails, revise before returning.
Example
See examples/input.md for a realistic discharge-summary snippet with multiple dosages, numeric thresholds, and clinical abbreviations. See examples/output.md for the complete patient-friendly rewrite, verification list, and rubric.
What ships with it: 4 files
16.1 KB alongside SKILL.md
.claude-plugin/
- plugin.json386 B
- README.md5.5 KB
Gives 0 of the 12 instructions most healthcare skills give in ~1.1k tokens
Counted across 147 of the 152 authors here whose files we hold, read 2026-08-07
- Export trial data to CSV formatin 11 of 147, across 2 files
- Retrieve trial details using an NCT IDin 11 of 147, across 2 files
- Split clinical datasets strictly by patientin 11 of 147, across 3 files
- Use the ClinicalTrials.gov API v2in 10 of 147, across 1 file
- Search trials by condition, drug, location, status or phasein 10 of 147, across 1 file
- Use maximum page size for bulk data retrievalin 10 of 147, across 1 file
- Extract and summarize key study informationin 10 of 147, across 1 file
- Combine multiple filters for targeted searchesin 10 of 147, across 1 file
- Print and review dataset statistics before modelingin 8 of 147, across 1 file
- Start model development with simple baselinesin 8 of 147, across 1 file
- Match preprocessing processors directly to data typesin 8 of 147, across 1 file
- Monitor validation metrics for task type and class imbalancein 8 of 147, across 1 file
Said here and by no other author read
- obtain clinical text and target audience
- prompt for missing text or target audience
- build an inventory of dosages and numeric values
- preserve every dosage and numeric value verbatim
- keep sentences under twenty words
- use plain equivalents instead of Latin phrases
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.