Dr. Avinash kumar gupta

← Back to README

VibeRounds Prompt Directory — Modules 1–9

Coined and developed by Dr. Avinash Kumar Gupta · Vibe Rounds, June 2026

Each prompt is tagged with one or more of four AI reasoning modes:

Two prompts carry dual tags where their design genuinely spans two modes.

Summary Statistics

AI Mode Prompt count (primary) % of all prompts % of all tag assignments
Traditional 59 62.1% 60.8%
Socratic 17 17.9% 17.5%
Brainstorming 11 11.6% 11.3%
Devil’s Advocate 10 10.5% 10.3%
Total prompts 95   97 tag assignments

Two prompts carry dual tags (1.1 = Socratic + Brainstorming; 9.0 = Traditional + Socratic), which is why tag assignments (97) exceed prompt count (95).

Per-Module Breakdown

Module Total Traditional Socratic Devil’s Advocate Brainstorming
Module 1 — Socratic Clinical Reasoning 11 4 5 2 1
Module 2 — Patient-Advocate Case Documentation 9 8 1 0 0
Module 3 — Extended Patient-Advocate Monitoring 8 7 1 0 0
Module 4 — Peer-Level Ward Round Preparation 10 5 4 1 0
Module 5 — Real-Time Case Review & Data Audit 12 9 3 0 0
Module 6 — Registry-Level Analytics 14 8 0 2 4
Module 7 — Longitudinal & Cross-Case Learning 13 6 1 2 4
Module 8 — Socratic-Mode Design Specification 5 3 1 1 0
Module 9 — N-of-1 Case Research Protocol 13 9 1 2 2
Total 95 59 17 10 11

Key pattern: Modules 2, 3, and 5 are almost entirely Traditional (documentation and audit workflows). Module 1 is the Socratic core of the system. Modules 6 and 7 are the only modules with zero or near-zero Socratic prompts — registry analytics is a data-querying, not a pedagogy, context. Devil’s Advocate prompts are scattered intentionally at closure/debrief steps across all modules.

Full Prompt Index

ID Module Phase Prompt purpose AI mode tags
1.0 Module 1 Initiation Session setup — Socratic contract, answer-withholding, curiosity hook Socratic
1.1 Module 1 Execution Generate a long list of Socratic questions around the case Socratic · Brainstorming
1.2 Module 1 Execution Socratic teacher on interventions — one question per intervention Socratic
1.3 Module 1 Execution Knowledge-scoped examination guide (textbook-restricted) Traditional
1.4 Module 1 Execution Mid-session reasoning checkpoint — grade logic, prioritisation, uncertainty Traditional
1.5 Module 1 Execution Bloom’s Taxonomy progression — six levels of clinical reasoning Socratic
1.6 Module 1 Execution Fink FLINK engagement — six significant-learning dimensions Socratic
1.7 Module 1 Closure End-of-case teaching summary — pearls, strengths, reading rec Traditional
1.8 Module 1 Closure Missed diagnosis debrief — cognitive bias, growth framing Devil’s Advocate
1.9 Module 1 Closure Difficulty ratchet — next session level and practice task Traditional
1.10 Module 1 Closure Critical awareness debrief — biases, risks, unresolved uncertainty Devil’s Advocate
2.0 Module 2 Initiation Case opening — compassionate documentation companion onboarding Traditional
2.1 Module 2 Execution Symptom & classifier capture by narrative method Traditional
2.2 Module 2 Execution Clinical examination guidance in plain language for advocate Traditional
2.3 Module 2 Execution Prescription transcription & dosage capture Traditional
2.4 Module 2 Execution SOAP note generation from case record Traditional
2.5 Module 2 Execution Data completeness audit — friendly gap checklist Traditional
2.6 Module 2 Closure Case record sign-off checklist Traditional
2.7 Module 2 Closure Advocate handover brief — 150-word read-aloud note Traditional
2.8 Module 2 Closure Bloom’s layer — advocate learning check (Remember → Apply) Socratic
3.0 Module 3 Initiation Baseline domain snapshot across four monitoring domains Traditional
3.1 Module 3 Execution Lifestyle monitoring — diet, sleep, mobility, fluid compliance Traditional
3.2 Module 3 Execution Mood & coping monitoring — emotional state, stressors, resilience Traditional
3.3 Module 3 Execution Medication monitoring & adherence — current vs discontinued table Traditional
3.4 Module 3 Execution Critical alert & red-flag triage — ALERT block generation Traditional
3.5 Module 3 Execution Fink FLINK monitoring reflection — six dimensions at checkpoint Socratic
3.6 Module 3 Closure Monthly domain review — Improved/Stable/Deteriorated per domain Traditional
3.7 Module 3 Closure Monitoring closure or escalation decision Traditional
4.0 Module 4 Initiation Study partner context load — peer framing, drug/investigation flags Traditional
4.1 Module 4 Execution Pre-rounds consultant prep — anticipate top 3 senior questions Socratic
4.2 Module 4 Execution Ward admission & pre-op checklist — monitoring, clearance, blood products Traditional
4.3 Module 4 Execution Cross-system complication & polypharmacy screen Devil’s Advocate
4.4 Module 4 Execution Night shift stat call triage — step-by-step deterioration reasoning Socratic
4.5 Module 4 Execution Bloom’s taxonomy round prep — six levels on most complex patient Socratic
4.6 Module 4 Execution Fink FLINK ward round reflection — six dimensions post-round Socratic
4.7 Module 4 Closure End-of-round patient status update — two-sentence progress notes Traditional
4.8 Module 4 Closure Ward round learning debrief — stakes, gap, reading rec Traditional
4.9 Module 4 Closure Formal SBAR handover generation Traditional
5.0 Module 5 Initiation Case context prime — confirm AI reading before analysis Traditional
5.1 Module 5 Execution Drug interaction & prescription audit with clinical learning angle Traditional
5.2 Module 5 Execution Recent interventions recap — 5 most recent + expected outcomes Traditional
5.3 Module 5 Execution Reformat raw log — cleaner layout + flag clinically significant datum Traditional
5.4 Module 5 Execution Two-week case refresh + reasoning question before AI reveals view Socratic
5.5 Module 5 Execution Patient journey summary — narrative arc from presentation to now Traditional
5.6 Module 5 Execution Masterclass follow-up — Socratic deep-dive on clinical principle Socratic
5.7 Module 5 Execution Patient voice — plain-language patient self-reflection prompt Traditional
5.8 Module 5 Execution Data anomaly flagging — impossible values, timestamps, contradictions Traditional
5.9 Module 5 Execution Fink FLINK deep case reflection — six dimensions Socratic
5.10 Module 5 Closure Case audit summary & action list — quality rating, concerns, gaps Traditional
5.11 Module 5 Closure Exportable case abstract — 200 words + Bloom’s level tag Traditional
6.0 Module 6 Initiation Registry health check — totals, completeness, known integrity issues Traditional
6.1 Module 6 Execution Level 1 — system scale & workforce yield; uncommon diagnosis lists Brainstorming
6.2 Module 6 Execution Level 2 — core operations & top diagnostic distribution Traditional
6.3 Module 6 Execution Level 3 — intervention mapping (pharmacological & procedural) Traditional
6.4 Module 6 Execution Level 4 — case-based learning: high-value N-of-1s, similar pairs, high-risk Brainstorming
6.5 Module 6 Execution Level 5 — network dynamics & workforce tier breakdown Traditional
6.6 Module 6 Execution Level 6 — diagnostic blindspots & pre-convergence phenotype detection Devil’s Advocate
6.7 Module 6 Execution Level 7 — granular ADR and polypharmacy auditing Traditional
6.8 Module 6 Execution Level 8 — advanced network triage: missing parameters, re-logging cases Traditional
6.9 Module 6 Execution Level 9 — data quality, narrative depth rating, divergent-outcome pairs Brainstorming
6.10 Module 6 Closure Analytics session synthesis — Registry Intelligence Report Traditional
6.11 Module 6 Closure Registry improvement wishlist — top 5 missing fields + ontology mapping Brainstorming
6.12 Module 6 Closure Exportable analytics summary — 300-word executive report Traditional
6.13 Module 6 Closure Critical awareness registry audit — selection bias, methodology critique Devil’s Advocate
7.0 Module 7 Initiation Learning inquiry setup — scope, time window, data completeness check Traditional
7.1 Module 7 Execution Longitudinal trajectory analytics — unresolved, fragmented, richest cases Traditional
7.2 Module 7 Execution Diagnostic evolution analytics — stable vs revised diagnoses, reasons Traditional
7.3 Module 7 Execution Registry learning analytics — symptom clusters preceding rare dx, cross-lingual Brainstorming
7.4 Module 7 Execution Diagnostic turning point detection — recurring trigger observations Traditional
7.5 Module 7 Execution Cross-case pattern mining — unexpected clusters, recurring constellations Brainstorming
7.6 Module 7 Execution Clinical reasoning analytics — cognitive bias patterns across registry Devil’s Advocate
7.7 Module 7 Execution Longitudinal case quality rating — five criteria, 1–10 per case Traditional
7.8 Module 7 Execution Bloom’s registry learning prompt — six levels at scale Socratic
7.9 Module 7 Closure Learning session synthesis — top patterns, blindspots, practice changes Traditional
7.10 Module 7 Closure New hypothesis generation — 3 testable N-of-1 hypotheses Brainstorming
7.11 Module 7 Closure Registry-as-curriculum builder — 5-case teaching sequence Brainstorming
7.12 Module 7 Closure Critical awareness cross-case debrief — artefacts, biostatistician challenge Devil’s Advocate
8.0 Module 8 Initiation Spec activation — review 12-point design criteria before authoring Socratic
8.1 Module 8 Closure Prompt peer review — score against 12-point spec, top 2 revisions Traditional
8.2 Module 8 Closure Regression test — simulate ‘idk’ for 3 turns, flag min-effort failures Devil’s Advocate
8.3 Module 8 Closure Difficulty calibration check — Year-1 vs senior resident side-by-side Traditional
8.4 Module 8 Closure Humanistic criteria audit — test criteria 11 & 12 with average-performing learner Traditional
9.0 Module 9 Initiation Single-session orchestration — all 7 stages in one prompt with pause points Traditional · Socratic
9.1 Module 9 Initiation Stage 1: index case structuring — extract diagnosis & intervention term lists Traditional
9.1a Module 9 Initiation Stage 1a: completeness check — second-pass scan for missed terms Traditional
9.2 Module 9 Execution Stage 2: comparator identification — search seeds, match criteria, source declaration Traditional
9.2a Module 9 Execution Stage 2a: PRISMA-style search-strategy note export Traditional
9.3 Module 9 Execution Stage 3: aggregation & cross-comparison — shared features vs divergence per criterion Traditional
9.4 Module 9 Execution Stage 4: two-tier analysis — coded (registry) view vs full narrative view, kept separate Devil’s Advocate
9.5 Module 9 Execution Stage 5: open-ended analytic ideation — expand option space, no pre-filtering Brainstorming
9.6 Module 9 Execution Stage 6: selection of analytic menu — restate chosen items as runnable tasks Brainstorming
9.6a Module 9 Execution Stage 6a: run an individual menu item (repeatable) Traditional
9.7 Module 9 Closure Stage 7: convergence — severity rating, priority rating, timeline, correlation table Traditional
9.8 Module 9 Closure Manuscript/CARE-field export — map outputs to CARE fields, flag gaps Traditional
9.9 Module 9 Closure Critical awareness debrief — bias, clinical risk, methodological critique Devil’s Advocate

Tagging rationale

Traditional was assigned when the AI’s primary role is to produce a structured deliverable (note, checklist, summary, table, report) and the user is a recipient, not a reasoning participant.

Socratic was assigned when the prompt explicitly withholds the answer, requires learner commitment before any hint, and advances only after the learner responds — the defining criteria from Module 8’s 12-point spec (criteria 1, 2, 10).

Devil’s Advocate was assigned when the prompt’s explicit goal is to surface what could be wrong, biased, or challenged — including cognitive bias audits, critical awareness debriefs, and stress-tests of the AI’s own outputs.

Brainstorming was assigned when the prompt’s explicit goal is expansion of the option space — generating lists, clusters, hypotheses, or analytic candidates without premature filtering or ranking.

VibeRounds Prompt Modules · Dr. Avinash Kumar Gupta · June 2026 ← Back to README