# Student Study OS — four governed study skills

This pack turns your Nurse AI OS into a study system built on how nursing actually teaches clinical judgment — and on what the evidence says AI tutoring should and shouldn't do.

**What's inside**

| File | What it does | Tier |
|---|---|---|
| `Socratic-Tutor.SKILL.md` | Questions, never answers — tutoring that builds your reasoning instead of replacing it | 🟢 Green |
| `NGN-Drill.SKILL.md` | Next Generation NCLEX–style practice items built on the NCSBN clinical-judgment steps, with rationales after you commit | 🟢 Green |
| `Sim-Case-NCJMM-Tanner.SKILL.md` | Fictional patient cases in gated phases, then a required Tanner-model debrief | 🟢 Green |
| `Flashcard-Forge.SKILL.md` | Your notes → active-recall cards on a spaced-repetition schedule | 🟢 Green |
| `Academic-Integrity-Header.md` | The shared boundary every skill above loads — including the graded-work refusal | — |
| `AI-Use-Disclosure-Template.md` | A ready disclosure block for assignments, generated from your session log | 🟡 Yellow |

**Setup (2 minutes).** Tell your AI: *"Load the four skills in 15-Student-Study-OS, including the Academic-Integrity-Header they share. My guidance tier is [novice / intermediate / advanced]. Schedule my flashcard review as a daily reminder."* Adjust your guidance tier any time — it controls how much hinting you get, not how much respect.

**The evidence behind the design (and its limits).**
- Socratic, guidance-tiered, phase-gated tutoring mirrors a published clinical-tutor architecture — but that work is a proof-of-concept without learner-outcome data yet: treat the *format* as promising, not proven.
- The NCJMM is NCSBN's *measurement* model for the licensure exam (it doesn't define clinical judgment); pairing it in-scenario with **Tanner's model for reflection** follows published teaching practice, and a controlled trial of Tanner-structured simulation improved knowledge and performance — while showing reflection is the step students skip, which is why the debrief here is mandatory, and why emotional care gets its own prompt.
- Practice analytics here are **readiness signals, not predictions**. Nothing in this pack predicts your NCLEX result, and confidence is tracked separately from mastery because the evidence shows they move separately.
- Integrity is built on **process transparency, not detection**: peer-reviewed piloting of AI-detection rubrics failed reliability, and detectors' false positives punish honest students. Your protection is the trail this pack generates for you.

**The five safety rules apply here fully** — above all: clinical-placement patients never enter this system. Practice on the fictional patients these skills generate, never on real ones.

> Agents propose. Humans judge. Nurses steward.