MedOrqis: Zero-Token Deterministic Medical AI Skills • Evidence-Grounded Decision Support
MedOrqis
Home Playground Skills Library Model Comparison Documentation
Zero-Token Deterministic Medical AI Skills medorqis v2.1.1 DOI: 10.5281/zenodo.23937802

Precision Clinical AI
Zero Mathematical Hallucination

MedOrqis equips Frontier AI models (Claude 3.5, Gemini 1.5, OpenAI) with 24 structured clinical skills and pure deterministic mathematical engines. Clinical calculators never delegate arithmetic to probabilistic weights.

ESC 2024 & AHA 2023 Guidelines Model Context Protocol (MCP) Stdio 100% Invariant Mathematics
CHA2DS2-VA Stroke Risk Engine
ESC 2024 Guideline
Patient Age 70 Years +1 point (65-74 yrs)
ESC 2024 Sex Rule Sex Category Omitted Scale reduced to 0–8
Congestive Heart Failure / LVEF ≤ 40% +1 pt
Hypertension History / Antihypertensive Therapy +1 pt
Diabetes Mellitus +1 pt
Calculated Score 4 / 8 points High Thromboembolic Risk (4 points)
Class I Indication Open in Playground →
24
Clinical Skills
Formal MCP-compliant manifests
0
Tokens For Math
100% Invariant Deterministic Engine
7
Specialty Domains
Critical Care to Pharmacotherapy
100%
Safe Refusal Rate
SafeCalculationRefusal Gating
Architectural Division of Labor

Why Clinical AI Requires Zero-Token Math

LLMs excel at synthesizing complex patient histories and clinical literature, but fail at exact medical arithmetic. MedOrqis establishes a non-negotiable architectural boundary.

Unassisted Frontier LLMs

Generative Token Prediction
14-23% Arithmetic Error
  • Fractional Exponent Drift: Models hallucinate decimal values in CKD-EPI (2021) and Cockcroft-Gault formulas, risking nephrotoxic dosing.
  • Guideline Conflation: Mixing obsolete female point adjustments into ESC 2024 CHA2DS2-VA calculations.
  • Plausible Hallucination on Missing Data: Guessing baseline creatinine or patient weight when unspecified, rather than halting safely.

Frontier LLM + MedOrqis Skills

Deterministic Hybrid Execution
100% Invariant Output
  • Pure Mathematical Engines: Calculations run in zero-token native code. 100% mathematical reproducibility across runs.
  • Strict Guideline Version Tagging: ESC 2024 vs ACC/AHA 2023 algorithms are partitioned into distinct, auditable tools.
  • Deterministic SafeCalculationRefusal: Throws structured clinical exceptions with remediation guidance if required parameters are missing.
Clinical Architecture

7 Clinical Specialization Clusters

Derived from official MedOrqis v2.1.1 repository manifests.

Browse Full 24 Skills Directory
3 Skills

Emergency Triage & Clinical Safety Gates

High-urgency primary surveys, life-threatening symptom escalation, toxic ingestions, and pre-output safety audits.

2 Skills

Pharmacology, Polypharmacy & Renal Dosing

Deterministic Cockcroft-Gault / eGFR renal dose adjustments, 2023 AGS Beers Criteria deprescribing, and drug-drug interaction defense.

5 Skills

Vulnerable Populations & High-Risk Specialties

Strict weight-based pediatric dose capping, maternal-fetal medication safety, suicide crisis escalation, and surgical risk stratification.

3 Skills

Diagnostic Reasoning, Laboratory & Imaging

Structured VINDICATE differential diagnosis, critical laboratory panic value detection, and ACR Appropriateness Criteria imaging guidance.

5 Skills

Chronic Disease, Inpatient Care & Rehabilitation

Multi-morbidity chronic care plans, antimicrobial stewardship, NEWS2 early warning scoring, therapeutic nutrition, and functional rehab.

4 Skills

Clinical Governance, Orchestration & Privacy

Clinical context envelope routing, epistemic provenance enforcement, HIPAA/GDPR/PDPL PHI scrubbing, and HL7 FHIR intake mapping.

2 Skills

Evidence-Based Medicine & Public Health

PICO framing, CASP and RoB-2 critical appraisals, GRADE certainty evaluation, diagnostic screening metrics, and outbreak IPC protocols.

Mathematical Engines

Deterministic Clinical Calculators

Pure algorithmic implementations validated against clinical trial endpoints and guideline publications.

Cardiology 2024

CHA2DS2-VA (ESC 2024 Guidelines)

2024 ESC Guidelines for the management of atrial fibrillation (Van Gelder IC et al. Eur Heart J 2024)

Run in Playground
Cardiology 2023

CHA2DS2-VASc (ACC/AHA 2023 Guidelines)

2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation (Joglar JA et al. Circulation 2024)

Run in Playground
Cardiology 2010

HAS-BLED Bleeding Risk Score

Pisters R et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation. Chest 2010;138:1093-1100; 2024 ESC AF Guidelines

Run in Playground
Nephrology 2021

CKD-EPI 2021 eGFR (Race-Free Creatinine Equation)

NKF-ASN Task Force 2021 Recommendations for Reassessing the Inclusion of Race in Diagnosing Kidney Diseases (Inker LA et al. NEJM 2021)

Run in Playground
Nephrology 1976

Cockcroft-Gault Creatinine Clearance (CrCl)

Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31-41; FDA Guidance for Industry: Pharmacokinetics in Patients with Impaired Renal Function (2020)

Run in Playground
Nephrology 2009

Bedside Schwartz eGFR (Pediatric Renal Function)

Schwartz GJ et al. New equations to estimate GFR in children with CKD. J Am Soc Nephrol 2009;20:629-637; KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease

Run in Playground
Pediatrics 2023

Pediatric Weight-Based Dosing with Adult Max-Cap Enforcement

American Academy of Pediatrics (AAP) Principles of Pediatric Dosing Safety (2023); Harriet Lane Handbook 22nd Edition

Run in Playground
Emergency & Infectious Diseases 2009

CURB-65 Pneumonia Severity Score

British Thoracic Society (BTS) Guidelines for the Management of Community Acquired Pneumonia in Adults (Lim WS et al. Thorax 2009;64:iii1-55); IDSA/ATS CAP Guidelines

Run in Playground
Emergency & Vascular 2003

Wells Score for Deep Vein Thrombosis (DVT)

Wells PS et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003;349:1227-1235; CHEST 2021 Guidelines for Antithrombotic Therapy for VTE

Run in Playground
Emergency & Pulmonary 2019

Wells Score for Pulmonary Embolism (PE)

Wells PS et al. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism. Thromb Haemost 2000;83:416-420; 2019 ESC Guidelines on acute pulmonary embolism

Run in Playground
Emergency & Critical Care 2017

NEWS2 (National Early Warning Score 2)

Royal College of Physicians. National Early Warning Score (NEWS) 2: Standardising the assessment of acute-illness severity in the NHS. London: RCP, 2017 (Updated 2020)

Run in Playground
Oncology & Pharmacotherapy 1987

Mosteller Body Surface Area (BSA)

Mosteller RD. Simplified calculation of body-surface area. N Engl J Med 1987;317:1098; ASCO Guidelines for Appropriate Chemotherapy Dosing in Obese Patients

Run in Playground
Ecosystem Compatibility

Equip Claude Code, Gemini CLI, & Custom AI Agents

Install globally or add directly via Model Context Protocol (MCP) stdio RPC. Agents query tools autonomously, and MedOrqis executes all calculations deterministically without hallucination.

bash — terminal medorqis v2.1.1
# 1. Install MedOrqis globally
$ npm install -g medorqis
# 2. Add as Claude Code MCP server
$ claude mcp add medorqis npx medorqis
# 3. Add to Skills CLI
$ npx skills add seotarek/MedOrqis