Each card explains what the skill does. Click Read it first to see every instruction inside it, then Download the ones you want. New to skills? Start with the guide.
Type or paste this into any Claude chat. Claude lists the skills you have switched on and what each one does. For a one-page list of all twenty skills and exactly what to type, open the cheat sheet and bookmark it.
Attach a paper and get a one-page brief: the clinical question, the design, the main result with absolute numbers and confidence intervals, the three limitations that matter most, and a bottom line no stronger than the data. It tells you when it only had the abstract to work from.
Works through a study domain by domain: comparator, selection, confounding, outcomes, statistics and spin. You get a table rating how much each problem could change the conclusion, plus the one question the authors most need to answer.
Explains what a number in a paper really means, from relative and absolute risk to NNT, odds ratios, confidence intervals and predictive values. It shows the arithmetic, flags common misreadings, and gives you one sentence you could say to a patient.
Puts ACOG, SMFM, RCOG, NICE, WHO and other recommendations on one question side by side, each with its year and strength of evidence. It shows where they agree, where they differ and why, and never quotes a guideline from memory without saying so.
Paste an answer from any AI tool and it is checked claim by claim. Every statement is marked verified, incorrect, unverifiable or opinion, every citation is looked up, and you get an overall verdict and the single highest-risk error.
Confirms that each reference exists, corrects wrong details, and checks whether it actually supports the sentence it is attached to. Missing DOIs or pages are marked unconfirmed rather than filled in, so nothing invented reaches your manuscript.
Cuts a manuscript, abstract or section to a word limit without touching numbers, citations or the strength of any claim. It reports the new word count and lists any cut you need to approve.
Turns a focused finding into a research letter of about 600 words with the standard four parts and one key table or figure. It confirms the main finding with you first and leaves clear placeholders instead of inventing references.
Writes a one-page cover letter to the editor with the question, the main finding, why it suits the journal, and the statements journals require. Plain and confident, with no exaggeration of novelty.
Splits a decision letter into numbered comments and drafts a courteous, point-by-point reply stating exactly what changed and where. Comments that need your decision are listed at the top.
Helps you organize a fair and specific review: a summary, major and minor comments with suggested fixes, confidential notes to the editor, and a closing Recommendation. It reminds you to check the journal's policy on AI use in peer review.
Reviews clinical text for imprecise, outdated or stigmatizing terms such as "elective" induction, "provider", "failed" labor or "noncompliant". Each one comes with a precise, respectful alternative and the reason.
Measures the reading level of a consent form or patient document and rewrites it in plain language, aiming for grade 6 to 8. A checklist confirms that every required element, from risks to the right to withdraw, is still there.
Builds a one-page comparison of the options with benefits and harms in absolute numbers on the same denominator, plus questions about what matters to the patient. When the evidence clearly favors one option, it says so, because shared decision-making is not neutrality.
Writes a calm, plain-language handout on a condition, test or procedure, with clear steps, warning signs and when to call. It ends with a teach-back question and a dated source line.
Helps you plan a hard conversation, such as bad news, a poor prognosis or disclosing an error, using the SPIKES framework. You get the core message in plain words, phrases to use and avoid, and the questions you are likely to be asked.
Organizes an ethically difficult case using the professional responsibility model: beneficence-based and autonomy-based obligations, justice, and preventive ethics. It supports discussion and does not replace your institution's ethics consultation service.
Creates a fully fictional, staged teaching case for students or residents, with discussion questions, evidence-based teaching points and optional multiple-choice questions. It never draws on real patient charts.
Turns a study into an accurate LinkedIn post under 1,200 characters: the finding stated plainly, the key number, the main limitation and the citation. No hype, and it reports the character count.
Answers a health myth calmly and respectfully, leading with the accurate fact rather than the myth, with a named source and a practical takeaway. Length fits the channel, from a short reply to a one-page explainer.
The step-by-step guide shows the one setting to switch on, how to add a skill, how to check it worked, and what to do if it does not.