Version: 2.3.0 — 22 Jun, 2026

Production Security Hardening, Concurrency Safety & Event Loop Stability (v2.3.0)

This release focuses on production-grade security, database concurrency, and runtime stability. We've replaced legacy HMAC authentication with PBKDF2, secured unauthenticated endpoints, eliminated event loop crashes under load, bounded memory growth, and added queue-level locking to prevent concurrent processing collisions.

---

🔐 1. Security & Auth Hardening

Documentation: Security Overview Wiki · IRIS RBAC Wiki · README.md Security Section

Implementation: Router.cls · Debug.cls · auth_utils.py

PBKDF2-SHA256 Password Hashing with Auto-Upgrade

• Replaced legacy HMAC-SHA256 static-salt hashes with PBKDF2-SHA256 (100,000 iterations, random 16-byte salt)
• Automatic on-the-fly migration: legacy HMAC hashes are detected by their prefix and upgraded to PBKDF2 on the next successful login — zero-touch transition
• Timing-safe comparison via hmac.compare_digest() prevents side-channel attacks

Revoked Superuser from UnknownUser

• Removed the %All superuser role from the CSP Gateway's UnknownUser account
• Replaced with fine-grained roles: %DB_INTEROP-CODE, %DB_INTEROP-DATA, %HS_DB_INTEROP, %DB_INTEROPX0001R, %DB_INTEROPX0001V, %HS_ServiceRole, %HS_Administrator
• Explicit schema-level SQL grants (SELECT, INSERT, UPDATE, DELETE) on HSFHIR_X0001_S, HSFHIR_X0001_R, HSFHIR_X0001_V, and ClaimAudit schemas
• Web app MatchRoles updated to include %DB_INTEROPX0001R:%DB_INTEROPX0001V

Secured Token Introspection (/auth/introspect)

• Now requires either Bearer JWT token or Basic Authentication credentials
• Unauthenticated requests return 401 {"active": false, "error": "Unauthorized"}

Admin-Gated LLM Settings

• Both GET /settings/llm and POST /settings/llm enforce Admin role check via RequireAuth("Admin")

---

⚡ 2. Python Backend Stability

Documentation: LLM Router Architecture Wiki · Orchestration - AI Hub Wiki

Implementation: agent_orchestrator.py · llm_router.py

Event Loop Lifecycle Fix

• Replaced custom asyncio event loop management inside multi-threaded Embedded Python contexts with clean asyncio.run() invocations
• Eliminates RuntimeError: This event loop is already running crashes under concurrent load

Bounded LRU LLM Response Cache

• Replaced the unbounded in-memory cache dictionary with a bounded OrderedDict LRU cache (max 500 entries)
• LRU eviction: oldest entry popped when limit exceeded; cache hits move entry to end for recency preservation
• Prevents unbounded memory growth under sustained production load

---

🗄️ 3. Database & Queue Concurrency Safety

Documentation: Data Classes Wiki · README.md Queue Section

Implementation: Queue.cls · Engine.cls

Atomic Queue Operations

• Enqueue() wrapped in TSTART/TCOMMIT with node-level locks (^ClaimAuditAI("QueueLock", claimResponseId), 10s timeout) preventing concurrent inserts
• WorkerLoop() wrapped in atomic transactions — failed adjudication triggers TROLLBACK, reverting status changes
• Process-level lock (^ClaimAuditAI("QueueProcessLock"), 5s timeout) prevents multiple workers from processing the same queue item concurrently

Dead Letter Queue Enhancements

• Items with exhausted retries (default 3) automatically move to dead-letter status
• New MoveToDeadLetter(), RequeueFromDeadLetter() methods on Queue.cls
• New API routes: GET /system/dead-letter-queue, POST /system/dead-letter-queue/:id/requeue

Stats Cache

• After successful adjudication, aggregated statistics are written to ^ClaimAuditAI("Stats") for O(1) dashboard reads
• Eliminates the N+1 FHIR GET pattern on every dashboard load

Iteration & Seeding Reliability

• LoadSampleData() uses HS.FHIRServer.API.InteractionsStrategy.Delete() instead of raw SQL DELETE statements, wrapped in TSTART/TCOMMIT
• ClearAllData() uses the same strategy-based deletion for consistency
• Iteration properties use direct .propertyName access instead of .%Get("propertyName") for cleaner code

---

🖥️ 4. UI/Frontend Role Constraints

Documentation: UI Architecture Overview Wiki

Implementation: LLMSettings.tsx

Role-Gated LLM Settings View

• Integrates useRoleStore + PERMISSIONS from the frontend role checking system
• Non-Admin users see an "Access Restricted" placeholder view instead of the LLM configuration form
• Backend RequireAuth("Admin") on both GET and POST provides defense-in-depth

---

🧪 5. Test Suite Expansion

Documentation: Testing Overview Wiki · README.md Testing Section

Implementation: test_auth_utils.py · test_tier_orchestrator.py

• 108 total tests (up from 101) — all passing in 1m 13s
• NEW test_auth_utils.py: 100% coverage for PBKDF2 hashing, timing-safe verification, legacy HMAC detection, and auto-upgrade triggers
• NEW test_tier_orchestrator.py: Unit tests for parallel tier execution under ThreadPoolExecutor, configurable per-tier timeouts, and circuit breaker open/close state transitions
• Frontend: 18/18 Vitest tests passing
• E2E: 11/11 integration stages passing (seeding, adjudication, queue, escalation, approval, collusion, health, backup)

---

📦 6. DevOps & Dependencies

Implementation: requirements.txt · Dockerfile

• Pinned httpx==0.28.1 and mcp==1.27.2 in requirements.txt for reproducible builds
• CSP process cache flush commands registered to ensure live REST dispatch classes compile and reload cleanly

---

Full Changelog: c28e8aa...6eef7be

New Release
Application Name: HealthCoach AI – FHIR Practice Arena
Current Version: 1.0.3
New Version: 1.0.4
Release Date: 06/19/2026
Release Notes:
We are pleased to announce version 1.0.4 of HealthCoach AI – FHIR Practice Arena. This release is specifically designed to make the application fully testable on InterSystems Open Exchange without any external configuration or API keys.
We have addressed all previous feedback regarding testability. The application now runs out‑of‑the‑box with no mandatory setup.
What's New in This Release:

• The AI Coach is now fully functional even without a Groq API key. If no key is present, a helpful fallback response is provided, allowing reviewers to experience the full workflow without any configuration.
• Supabase integration is now optional and automatically detected. The app runs perfectly without it, using session‑based data.
• The Practice Problem tab is completely independent and works with any valid FHIR JSON or date input. No external services are required.
• A comprehensive Documentation tab has been added, clearly explaining prerequisites, how to start, how to test, and a live demo link.
• The video introduction is embedded and plays directly in the browser.
• All text and labels are fully translated into English, Français, and Español.
• The sidebar now includes a Supabase status indicator and a clear "How to use" guide.
  Key Features:
• Video introduction explaining the platform and its learning approach.
• Hands‑on practice problem: calculate a patient's age from a FHIR Patient resource.
• AI coach that provides instant, constructive feedback (fallback mode available if no Groq key).
• Optional Supabase logging for tracking user interactions.
• Multi‑language support.
• Complete documentation included.
  Live Demo:
  https://healthcoach-ai-fhir-training-uzdeefhyupvfjkywurau6w.streamlit.app/
  No login or setup is required – the app is fully open for testing.
  Repository:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training
  Technologies Used:
• Python 3.8+
• Streamlit
• Groq API (optional)
• Supabase (optional)
• FHIR (HL7)
  Feedback and Support:
  Please report issues or suggest features through GitHub Issues:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training/issues
  About the Developer:
  This application was built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise.
  Contact:
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  We are confident that this release will allow InterSystems Open Exchange reviewers to test the application instantly and without any barriers. We welcome your feedback and look forward to your evaluation.

Hospital Management System Software – Release Note
Version: 1.0.3
Release Date: June 19, 2026
We are pleased to announce a new release of the Hospital Management System Software, a comprehensive, multi‑specialty hospital management solution built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py.
This release is specifically designed to address the feedback that the application was “impossible to test” without external configuration. We have made the app fully self‑contained and ready for immediate demonstration.
What’s New in This Release

• The application now includes a complete set of embedded hospital guidelines (Patient Rights, Admission/Discharge, Medication, Infection Control, Emergency, Data Privacy, Code of Conduct, Patient Education, Quality). No external document upload is required.
• Default patient data (Emily Clark, James Brown, Sophia Lee) is automatically seeded, so the AI Diagnostic Assistant can answer patient‑specific questions without any prior setup.
• The AI Diagnostic Assistant now works even without a Groq API key – a helpful fallback response is provided, so reviewers can test the full workflow immediately.
• The guidelines are now trimmed to 2000 characters, allowing the AI to access the full Discharge Criteria and other later sections.
• Supabase integration is optional – the app runs perfectly without it, using session‑based data.
• The login credentials are clearly shown on the login page for easy access.
  Key Features
• Patient registration with unique Medical Record Number and full EMR.
• Billing and revenue cycle management.
• Pharmacy management with stock alerts and prescription dispensing.
• Laboratory test ordering and results tracking.
• Radiology and imaging scheduling.
• Inventory and financial management.
• Enterprise‑grade reporting dashboard.
• AI Diagnostic Assistant that answers clinical and operational questions based on actual hospital data and embedded guidelines.
• AI voice explanation for guided walkthroughs.
• Multi‑language support (English, Spanish, French).
  Live Demo
  The application is now fully testable without any installation or API keys.
  https://hospital-management-system-software-built-by-gesner-deslandes.streamlit.app/
  Login credentials:
  Username: Gesner
  Password: 20082010
  Role: Admin
  Repository
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes
  Technologies Used
• Python 3.10+
• Streamlit
• Groq API (optional)
• Supabase (optional)
• gTTS for voice
• Plotly for charts
• Pandas / NumPy
  License
  MIT License
  About the Developer
  Built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise – in healthcare, education, AI, and automation.
  Contact
  Email: deslandes78@gmail.com
  Phone / WhatsApp: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  Support
  For bug reports or feature requests, please open an issue on GitHub:
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes/issues
  We are confident that this release will allow InterSystems Open Exchange reviewers to test the application instantly and without any barriers. We welcome your feedback and look forward to your evaluation.

Version: 2.2.0

🚀 Payment Integrity Hardening & Resilient ML Architecture (v2.2.0)

We have released a major reliability and capability update for ClaimAuditAI. This release hardens the InterSystems IRIS database layer, expands the PyTorch autoencoder feature space, introduces robust ML backward compatibility, prevents offline unit test hangs, and increases unit test coverage.

---

🧠 1. Resilient 8-Dimensional ML Autoencoder & Backward Compatibility

• Documentation: [Autoencoder Architecture Wiki](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/05-ml/Autoencoder%20Architecture.md) & [README.md ML Section](file:///Users/mck/Desktop/claimauditai/README.md#2-the-3-tier-ai-engine-tier_orchestratorpy)
• Implementation: [autoencoder_train.py](file:///Users/mck/Desktop/claimauditai/src/python/autoencoder_train.py) & [tier_orchestrator.py](file:///Users/mck/Desktop/claimauditai/src/python/tier_orchestrator.py)
• Details:
  • Expanded Architecture: Scaled the unsupervised PyTorch Autoencoder from 5 to 8 input dimensions (introducing CodeCount, ServiceMonth, and ProviderBusyness columns) and increased the latent bottleneck to 6 dimensions to capture complex billing features.
  • Dynamic Dimension Detection: Resolved state_dict size mismatch load crashes when starting up with legacy models. The loader now dynamically extracts the saved model's dimensions from the means array inside stats.npz and instantiates the matching architecture.
  • Feature Slicing & Padding: Evaluated claim features are automatically sliced or padded to fit the loaded model's dimensions, ensuring seamless, hot-reloadable backward compatibility.
  • Weight Determinism: Enforces PyTorch seed initialization (torch.manual_seed(42)) immediately before model instantiation to eliminate non-deterministic loss variance.

---

🛠️ 2. IRIS Database Layer & Container Stability

• Documentation: [Python Validation and ML Model Pitfalls Wiki](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/09-troubleshooting/Python%20Validation%20and%20ML%20Model%20Pitfalls.md)
• Implementation: [Engine.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/AI/Engine.cls), [Interactions.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/FHIR/Interactions.cls), [Router.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/REST/Router.cls), [requirements.txt](file:///Users/mck/Desktop/claimauditai/requirements.txt)
• Details:
  • SQL Upsert Refactor: Replaced unsupported INSERT OR UPDATE SQL statements in Engine.cls with a high-performance check-then-act UPDATE and fallback INSERT pattern, resolving compile-time database errors.
  • ObjectScript Operator Fixes: Fixed incorrect use of the MUMPS logical OR (!) operator, replacing it with the standard short-circuiting logical OR (||) across multiple router, authentication, and engine classes.
  • Dependency Conflict Resolution: Pinned scipy==1.17.1 in the Dockerfile and requirements.txt to eliminate runtime attribute conflicts (AttributeError: module 'numpy' has no attribute 'long') between sentence-transformers, SciPy, and NumPy.

---

🧪 3. Test Suite Expansion & Offline Mocking Safeguards

• Documentation: [Testing Overview Wiki](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/08-testing/Testing%20Overview.md) & [README.md Testing Section](file:///Users/mck/Desktop/claimauditai/README.md#-testing--verification)
• Implementation: [test_queue.py](file:///Users/mck/Desktop/claimauditai/src/python/tests/test_queue.py), [test_dx_procedure_validator.py](file:///Users/mck/Desktop/claimauditai/src/python/tests/test_dx_procedure_validator.py), [test_llm_router.py](file:///Users/mck/Desktop/claimauditai/src/python/tests/test_llm_router.py)
• Details:
  • 101 Total Tests: Expanded test coverage from 92 to 101 test cases covering both database, validation, and ML models.
  • Embedded Python Queue Tests: Introduced test_queue.py utilizing native iris Python bindings to test enqueuing, moving to dead-letter, and requeueing database operations.
  • Offline Test Performance: Integrated global autouse mocks in test_llm_router.py and test_agent_orchestrator.py to prevent 310-second connection timeouts when running tests in environments where Ollama is unreachable. This reduced unit test execution time to 1m 13s with zero hangs.

---

📋 4. Terminology Validation Rule Hardening

• Documentation: [Terminology MCP Server Wiki](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/07-api/Terminology%20MCP%20Server.md)
• Implementation: [dx_procedure_validator.py](file:///Users/mck/Desktop/claimauditai/src/python/dx_procedure_validator.py)
• Details:
  • Alphanumeric Rule Resolution: Fixed unquoted validation rules (e.g. H0001 ranges) that caused runtime NameError exceptions during module import.
  • Type Mismatch Safety: Added dynamic casting of range boundaries (integer vs string) in the validation logic to prevent comparison TypeErrors during upcoding validation.

New Release
Application Name: HealthCoach AI – FHIR Practice Arena
Current Version: 1.0.2
New Version: 1.0.3
Release Date: 06/19/2026
Release Notes:
We are pleased to announce version 1.0.3 of HealthCoach AI – FHIR Practice Arena, an interactive educational platform designed to help healthcare IT professionals, developers, and students master FHIR interoperability with the assistance of an AI coach.
This release focuses on making the application fully testable and accessible for InterSystems Open Exchange reviewers and end users. The app now runs without requiring any external API keys, ensuring a smooth first-use experience.
What's New in This Release:

• The Groq API key is now optional. If no key is provided, the AI Coach falls back to a helpful generic response, so users can test the full workflow without any setup.
• Supabase integration is now optional and flexible. The app automatically detects Supabase credentials whether they are stored as flat keys or nested under a supabase section.
• Added a Supabase status indicator in the sidebar to show whether the optional logging feature is active.
• Improved error handling throughout the application – the app no longer crashes when secrets are missing.
• Updated the Documentation tab to reflect that the Groq key is optional, with clear instructions for both flat and nested secrets formats.
• The Practice Problem tab remains fully functional and independent of any API keys.
• The AI Coach tab now clearly indicates when it is running in fallback mode, so users understand the expected behavior.
  Key Features:
• Video introduction explaining the platform and its learning approach.
• Hands-on practice problem: calculate a patient's age from a FHIR Patient resource.
• AI coach powered by Groq's Llama 3.1 (optional – fallback mode available).
• Optional Supabase logging for tracking user interactions.
• Multi-language support for English, Français, and Español.
• Comprehensive documentation included.
  Live Demo:
  https://healthcoach-ai-fhir-training-uzdeefhyupvfjkywurau6w.streamlit.app/
  Repository:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training
  Technologies Used:
• Python 3.8+
• Streamlit
• Groq API (optional)
• Supabase (optional)
• FHIR (HL7)
  Feedback and Support:
  Please report issues or suggest features through GitHub Issues:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training/issues
  About the Developer:
  This application was built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise.
  Contact:
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  We are committed to continuous improvement and welcome collaboration opportunities. Explore the demo, test the features, and let us know how we can serve your learning needs.

• fix indentation in docker-compose.yml
• fix some typos + URL in Read.Me
• remove unnecessary webterminal / iterm

Initial Release – System Health AI Monitor
Version: 1.0.0
Release Date: June 19, 2026
We are pleased to announce the initial release of the System Health AI Monitor, a powerful real‑time system observability dashboard with AI‑powered anomaly detection and predictive analytics.
This release introduces a comprehensive set of features to help platform engineers, software architects, and IT professionals monitor system health, detect anomalies, and gain actionable insights.
Key Features

• Real‑time monitoring of CPU, memory, disk usage, and network latency.
• Demo mode with simulated data for testing and demonstration.
• Live mode that reads actual system metrics using the psutil library.
• AI predictive analysis powered by Groq's Llama 3.1 model.
• Automatic anomaly detection with real‑time alerts for high CPU or latency.
• Detailed system information display: device type, brand, model, hostname, operating system, processor, memory, IP addresses, MAC addresses, network adapters, default gateway, DNS servers, and public IP.
• Automatic device type detection (Laptop / Desktop).
• Global Shield API key integration for security status display.
• AI voice explanation in English, French, and Spanish (female voice).
• Full multi‑language support for the interface.
• Step‑by‑step local installation instructions in the sidebar.
• Configurable auto‑refresh rate.
• Interactive Plotly charts for trend visualization.
  Live Demo
  You can test the fully functional application at the following link:
  https://system-health-ai-monitor-important-9bemdyosmbfmtx4t8wygbv.streamlit.app/
  No login required – just open the link and start testing.
  Repository
  https://github.com/Deslandes1/system-health-ai-monitor-Important
  Technologies Used
• Python 3.10+
• Streamlit
• Groq API (Llama 3.1)
• psutil
• netifaces
• edge-tts
• Plotly
• Pandas / NumPy
  License
  This project is licensed under the MIT License.
  About the Developer
  This software was built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise – in healthcare, education, AI, and automation.
  Contact and Pricing
  Email: deslandes78@gmail.com
  Phone / WhatsApp: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  Full license starting at 499 USD, including complete source code and setup support.
  Support
  For bug reports, feature requests, or any questions, please open an issue on GitHub:
  https://github.com/Deslandes1/system-health-ai-monitor-Important/issues
  We aim to respond within 48 hours.
  Built with ❤️ in Haiti by Gesner Deslandes – GlobalInternet.py

New Release
Application Name: AI Customer Service Suite
Current Version: 1.0.0
New Version: 1.0.0
Release Date: 06/18/2026
Release Notes:
We are pleased to announce the first release of AI Customer Service Suite, a powerful automation tool designed to help businesses manage customer support across text messages, emails, and phone/WhatsApp calls.
This release includes the following features:

• AI-powered responses based on your company guidelines
• Multi-language support in English, French, and Spanish
• Email auto-reply via Gmail with processing of up to 100 emails per day
• Twilio integration for voice and WhatsApp calls
• Embedded company guidelines with no file upload needed
• AI female voice description for social media and demos
• Light purple theme with modern user interface
• Email processing log with date and time stamps
• Secure API connections with Groq
  Live Demo:
  https://ai-customer-service-suite-bemey6yywchvkz7yrghufc.streamlit.app/
  Repository:
  https://github.com/Deslandes1/ai-customer-service-suite
  Technologies Used:
• Python
• Streamlit
• Groq API
• Edge TTS for AI voice
• Gmail API for IMAP and SMTP
• PyPDF2 and python-docx for file processing
  Feedback and Support:
  Please report issues or suggest features through GitHub Issues:
  https://github.com/Deslandes1/ai-customer-service-suite/issues
  About the Developer:
  This application was built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise.
  Contact:
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  This is the first release. Future updates will include additional features and improvements based on user feedback.

New Release

New Release
Application Name: HealthCoach AI – FHIR Practice Arena
Current Version: 1.0.1
New Version: 1.0.2
Release Date: 06/18/2026
Release Notes:
We are pleased to announce version 1.0.2 of HealthCoach AI – FHIR Practice Arena, an interactive educational platform designed to help healthcare IT professionals, developers, and students master FHIR interoperability with the assistance of an AI coach.
This release addresses feedback from the InterSystems Open Exchange review and significantly improves documentation, usability, and functionality.
What's New in This Release:

• Added a complete Documentation tab with clear sections for prerequisites, how to start, how to test, and a live demo link.
• Enhanced the AI Coach with improved feedback on code correctness, edge cases, and FHIR best practices. The AI now provides more detailed and constructive suggestions.
• Updated the sidebar with developer information, contact details, and a step-by-step usage guide.
• Improved multi-language support for English, Français, and Español for all interface elements.
• Made the application fully launchable with clear setup instructions. The live demo is now accessible without any installation.
• Fixed all known issues related to deployment and testing.
  Key Features:
• Video introduction explaining the platform and its learning approach
• Hands-on practice problem: calculate a patient's age from a FHIR Patient resource
• AI coach powered by Groq's LLM that provides instant, constructive feedback
• Secure, encrypted environment for all interactions
• Multi-language interface
• Comprehensive documentation included
  Live Demo:
  https://healthcoach-ai-fhir-training-uzdeefhyupvfjkywurau6w.streamlit.app/
  Repository:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training
  Technologies Used:
• Python 3.8+
• Streamlit
• Groq API
• FHIR (HL7)
  Feedback and Support:
  Please report issues or suggest features through GitHub Issues:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training/issues
  About the Developer:
  This application was built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise.
  Contact:
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  We are committed to continuous improvement and welcome collaboration opportunities. Explore the demo, test the features, and let us know how we can serve your learning needs.

Applications
Releases
New Release
Application Name: Hospital Management System Software
Current Version: 1.0.1
New Version: 1.0.2
Release Date: 06/18/2026
Release Notes:
We are pleased to announce version 1.0.2 of the Hospital Management System Software, a comprehensive, multi-specialty hospital management solution built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. This release includes significant improvements to stability, functionality, and user experience.
What's New in This Release:

• Supabase Integration – patient records, invoices, and lab orders now persist across sessions. Data is stored securely in the cloud, making the application ready for real-world use.
• AI Diagnostic Assistant – upgraded with a robust timeout mechanism and improved error handling. The AI now provides reliable answers based on actual hospital data and embedded guidelines.
• Embedded Hospital Guidelines – the app includes a comprehensive set of standard hospital guidelines covering Patient Rights, Admission and Discharge, Medication Management, Infection Control, Emergency Procedures, Data Privacy, Code of Conduct, Patient Education, and Quality Improvement.
• AI Voice Explanation – a female AI voice describes the entire system on demand. Click the button in the sidebar for a guided walkthrough of all modules.
• Multi-Language Support – the interface is now available in English, Spanish, and French.
• Improved Error Handling – all user actions return clear feedback. The app no longer hangs, crashes, or times out.
• Modern User Interface – a clean, light blue design with intuitive navigation and responsive layout.
• Full Enterprise Reporting – generate reports on daily revenue, patient visits, pharmacy sales, and department performance.
  Key Features:
• Patient registration with unique Medical Record Number and full Electronic Medical Records
• Billing and revenue cycle management with invoice generation
• Pharmacy management with stock alerts and prescription dispensing
• Laboratory integration with test ordering and results tracking
• Radiology and imaging scheduling
• Inventory and financial management
• Enterprise-grade reporting dashboard
• AI Diagnostic Assistant for clinical and operational questions
• AI voice explanation for guided software walkthroughs
• Secure login with role-based access control
  Live Demo:
  https://hospital-management-system-software-built-by-gesner-deslandes.streamlit.app/
  Login Credentials:
  Username: Gesner
  Password: 20082010
  Role: Admin
  Repository:
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes
  Technologies Used:
• Python
• Streamlit
• Groq API
• Supabase for cloud data persistence
• gTTS for AI voice generation
• Python-docx for document processing
• Fully responsive and mobile-friendly
  Feedback and Support:
  Please report issues or suggest features through GitHub Issues:
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes/issues
  About the Developer:
  This application was built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise in healthcare, education, and beyond.
  Contact:
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  We are committed to continuous improvement and welcome collaboration opportunities. Explore the demo, test the features, and let us know how we can serve your hospital's needs.

Hospital Management System Software – Release Note v2.0
We are pleased to announce a new release of the Hospital Management System Software, a comprehensive, multi‑specialty hospital management solution built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py. This release marks a major milestone, with the application now fully functional, stable, and ready for demonstration.
What's New in This Release

• Supabase Integration – patient records, invoices, and lab orders now persist across sessions. Data is stored securely in the cloud, making the application ready for real‑world use.
• AI Diagnostic Assistant – upgraded with a robust timeout mechanism and improved error handling. The AI now provides reliable answers based on actual hospital data and embedded guidelines.
• Embedded Hospital Guidelines – the app includes a comprehensive set of standard hospital guidelines covering Patient Rights, Admission and Discharge, Medication Management, Infection Control, Emergency Procedures, Data Privacy, Code of Conduct, Patient Education, and Quality Improvement.
• AI Voice Explanation – a female AI voice describes the entire system on demand. Click the button in the sidebar for a guided walkthrough of all modules.
• Multi‑Language Support – the interface is now available in English, Spanish, and French.
• Improved Error Handling – all user actions return clear feedback. The app no longer hangs, crashes, or times out.
• Modern User Interface – a clean, light blue design with intuitive navigation and responsive layout.
• Full Enterprise Reporting – generate reports on daily revenue, patient visits, pharmacy sales, and department performance.
  Key Features
• Patient registration with unique Medical Record Number and full Electronic Medical Records
• Billing and revenue cycle management with invoice generation
• Pharmacy management with stock alerts and prescription dispensing
• Laboratory integration with test ordering and results tracking
• Radiology and imaging scheduling
• Inventory and financial management
• Enterprise‑grade reporting dashboard
• AI Diagnostic Assistant for clinical and operational questions
• AI voice explanation for guided software walkthroughs
• Secure login with role‑based access control
  Technical Highlights
• Built with Python and Streamlit
• Powered by Groq's LLM for AI features
• Supabase for cloud data persistence
• gTTS for AI voice generation
• Fully responsive and mobile‑friendly
  Live Demo
  You can test the fully functional application at the following link:
  https://hospital-management-system-software-built-by-gesner-deslandes.streamlit.app/
  Login credentials:
  Username: Gesner
  Password: 20082010
  Role: Admin
  No installation is required – the demo runs in any browser.
  Repository
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes
  Feedback and Support
  We welcome your feedback. Please report issues or suggest features through our GitHub Issues page:
  https://github.com/Deslandes1/Hospital-Management-System-Software-built-by-Gesner-Deslandes/issues
  About the Developer
  This application was built by Gesner Deslandes, Engineer‑in‑Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise – in healthcare, education, and beyond.
  Contact
  Email: deslandes78@gmail.com
  Phone: (509) 4738 5663
  Website: https://globalinternetsitepy-abh7v6tnmskxxnuplrdcgk.streamlit.app/
  GitHub: https://github.com/Deslandes1
  We are committed to continuous improvement and welcome collaboration opportunities. Explore the demo, test the features, and let us know how we can serve your hospital's needs.

Here is the clean release note with the "Next Steps" section removed, as requested:

HealthCoach AI – FHIR Practice Arena
Release Note – Version 1.1.0
We are pleased to announce a new release of HealthCoach AI – FHIR Practice Arena, an interactive educational platform designed to help healthcare IT professionals, developers, and students master FHIR interoperability with the assistance of an AI coach.
What's New in This Release

• Added a complete Documentation tab with clear sections for prerequisites, how to start, how to test, and a live demo link.
• Enhanced the AI Coach with improved feedback on code correctness, edge cases, and FHIR best practices.
• Updated the sidebar with developer information, contact details, and a step-by-step usage guide.
• Improved multi-language support for English, Français, and Español for all interface elements.
• Made the application fully launchable with clear setup instructions.
  Key Features
• Video introduction explaining the platform and its learning approach.
• Hands-on practice problem: calculate a patient's age from a FHIR Patient resource.
• AI coach powered by Groq's LLM that provides instant, constructive feedback.
• Secure, encrypted environment for all interactions.
  Live Demo
  https://healthcoach-ai-fhir-training-uzdeefhyupvfjkywurau6w.streamlit.app/
  Repository
  https://github.com/Deslandes1/healthcoach-ai-fhir-training
  Technologies Used
• Python 3.8+
• Streamlit
• Groq API
• FHIR (HL7)
  Feedback and Support
  We welcome your feedback. Please report issues or suggest features through our GitHub Issues page:
  https://github.com/Deslandes1/healthcoach-ai-fhir-training/issues
  About the Developer
  This application was built by Gesner Deslandes, Engineer-in-Chief at GlobalInternet.py. We deliver tailored software solutions connecting global markets with local expertise.

added video url

added pt-br version of the article

Triage Park V2.0.0 Release Notes

Deployed: 2026-06-15

Highlights

• Transformed from a single triage app into a full multi-agent clinical platform
• New supervisor orchestrator routes and chains specialist agents
• Triage core is now agentic: bounded tool-using reason loop + self-critique verifier
• Adaptive, multilingual intake with dynamic follow-up questions

New specialist agents

• Orchestrator — routes/chains the specialists
• Copilot — read-only, grounded Q&A over one patient's record
• Cohort — population-level risk & care-gap aggregation
• Care plan — drafts a structured care plan from the record
• Gaps-in-care — deterministic preventive-care / monitoring checks
• Labs explainer — plain-language explanation of recent labs
• Follow-up — flags abnormal / out-of-range results
• NL→FHIR query — natural language to a validated, read-only FHIR search
• Summary — role-aware clinical summary of the record

Safety & reliability

• Deterministic safety screening — medication-interaction and allergy checks
• Risk scoring via IntegratedML — IRIS readmission/deterioration model
• Test suite + CI — 11 new test files (safety gates, red-flags, reasoning, risk, orchestrator, interview, gaps) wired into GitHub Actions

New platform UI

• New views: Agent Toolbox, Assistant, Cohort analytics, Explore (NL query)
• Tabbed clinician console
• Reworked case-detail screen (+398 lines) and worklist
• Expanded API client and types

Deployment & infra

• Production deployment: Caddy reverse proxy, automatic HTTPS via sslip.io, DEPLOY.md
• Security hardening: agent/IRIS/UI bind to 127.0.0.1 only; Caddy is the sole public service

added logo

added first article link

Live demo added to the README

Updated model access

Updated models

Updated agents integration with mimo model from xiaomi

Initial Release

wire up real LLM inference engine

• agents now powered by Cloudflare Workers AI

Fixes Applied per Review Feedback

📝 Documentation Updates & Release Details (v2.1.1)

We have updated the project's README.md and key pages of the claimauditai-wiki to document production-grade hardening, security fallbacks, and the newly integrated Model Context Protocol (MCP) server.

---

🚀 1. Model Context Protocol (MCP) Terminology Server

• Documentation: Terminology MCP Server Wiki & README.md Details
• Implementation: mcp_server.py
• Details: Documents the new FastMCP terminology resolver deployed inside the claimaudit-iris container. It enables the AI chat assistant to call standardized local clinical tools to prevent LLM hallucinations:
  • lookup_cpt_code(code: str) -> str: Translates 5-digit CPT codes to official procedure descriptions.
  • lookup_icd_code(code: str) -> str: Translates diagnosis codes, supporting prefix matches.
  • validate_diagnosis_procedure(icd_code: str, cpt_code: str) -> str: Runs clinical compatibility audits.

---

🤖 2. Machine Learning Drift Protection & Safeguards

• Documentation: Autoencoder Architecture Wiki & README.md ML section
• Implementation: autoencoder_train.py
• Details: Documents model training protections introduced to ensure stability in production environments:
  • Validation Split ($15%$): During retraining, $15%$ of historical data is reserved for evaluation, and training is executed on the remaining $85%$.
  • Drift Threshold ($15%$): The candidate model's validation loss must not degrade by more than $15%$ compared to the active baseline model, or the update is safely rejected.
  • Rolling Backups: Automatically rotates and retains the last 3 successful model weights and stats (_v1.pth/npz, _v2.pth/npz, _v3.pth/npz) for instant recovery.
  • Weight Determinism: Details the usage of torch.manual_seed(42) before model instantiation to eliminate non-deterministic sequential loss fluctuations.
  • Bypass Safety: Gracefully bypasses evaluation (returning flagged=False) when there are fewer than 5 historical claims instead of throwing errors.

---

🔑 3. Non-Blocking JWT Security Fallback (Safe Mode)

• Documentation: API Key Handling Wiki & README.md Security section
• Implementation: Auth.cls
• Details: To guarantee reviewers can inspect and access the system without configuration blocks, the application now boots in Non-Blocking JWT Safe Mode if JWT_SECRET is missing.
  • If missing, the system logs a warning to the ^ClaimAuditSecurityError global and falls back to a persistent GUID rather than causing fatal startup exceptions.

---

📊 4. Direct Stats Extraction & Transaction Queues

• Documentation: README.md Adjudication Queue section
• Implementation: Router.cls & Queue.cls
• Details:
  • Extension Parsing: Refactored the dashboard statistics logic to directly parse structured https://claimauditai.com/fhir/extension/tier-results JSON objects inside ClaimResponse instead of using brittle text substrings.
  • Background Processing: Documents the asynchronous queue (ClaimAudit.Data.Queue) processing engine, which allows immediate 202 Accepted ingestion responses for clients while executing AI pipelines in background worker threads.

---

🧪 5. Expanded Verification & Testing Suites

• Documentation: README.md Testing section
• Implementation: test_mcp_server.py
• Details:
  • Python unit test cases have been expanded to 92 tests, verifying the MCP endpoints, drift validations, backup rotations, and graph cycles.
  • All Vitest, Pytest, and end-to-end integration tests have been validated to run successfully.

---

📚 New Wiki Documentation & Accuracy Audit

Recent commits (b0ffe6d, 081bf4f, and feda00f) added 6 new architectural documents to the claimauditai-wiki and updated the README.md to establish complete alignment between the codebase and the documentation:

• Data Classes Wiki — Documents the persistent storage classes (ChatHistory with reserved keyword escaping, background Queue, Cytoscape-compatible GraphStore, and Debug utilities).
• UI Architecture Wiki — Details the frontend React 19 + TypeScript + Vite 6 tech stack, Zustand 5 state stores, and TanStack Query 5 data-fetching structures.
• LLM Router Architecture Wiki — Explains the failover queues, exponential backoff retries, sliding window rate limits, caching TTL, and SSE chat streaming in the backend.
• Agent Tool Registry Wiki — Covers the decorator-based function introspection schema generation for tools mapped to the ReAct agent loop.
• Diagnosis-Procedure Validator Wiki — Details rule-based CPT and ICD-10 compatibility audits and validation chapter range mappings.
• FHIR Extension Definitions Wiki — Documents custom billing, decision, and risk scoring extensions attached to pended ClaimResponse resources.

ClaimAuditAI — Release Notes v2.1.0 (Production Hardening & Drift Safeguards)

Welcome to the v2.1.0 Release of ClaimAuditAI. This release focuses on production-grade hardening, model drift protection, training determinism, and robust integrations with the Model Context Protocol (MCP) terminology server and structured FHIR extensions.

---

🛠 Architectural & Safety Enhancements

1. 🤖 PyTorch Autoencoder Drift Checking & Validation Split

To prevent model degradation in production, we introduced a robust verification loop during the autoencoder retraining process:

• Validation Split (15%): Retraining now reserves 15% of historical claims as a validation dataset, with the remaining 85% used for training.
• Validation Drift Safeguards (15% Threshold): Prior to overwriting the active model weights (autoencoder_model.pth), the candidate model is evaluated on the validation dataset. If the candidate loss represents a degradation of more than 15% compared to the baseline model's loss, the update is rejected, a warning is logged, and the previous model remains active.
• Rolling Backups (3 Versions): The system automatically rotates and maintains the last 3 successful model weights and stats (_v1.pth/npz, _v2.pth/npz, _v3.pth/npz) for instant rollback capability.
• Deterministic Weight Initialization: Resetting the seed (torch.manual_seed(42)) immediately before candidate model instantiation guarantees deterministic initialization, preventing false drift rejections in automated sequential test runners.

2. 🔑 Non-Blocking JWT Secret Review Fallback

• Location: [Auth.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/REST/Auth.cls)
• Hardening: Modified GetJWTSecret() to not block application startup or throw a fatal exception if CLAIMAUDIT_ENV=production is set but the JWT_SECRET environment variable is missing.
• Auditing: Logs a clear warning to the ^ClaimAuditSecurityError global and falls back to a persistent GUID. This ensures the application remains functional for contest reviewers while keeping a clear audit trail of security configuration gaps.

3. 📊 Direct Stats Parsing via FHIR Extensions

• Location: [Router.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/REST/Router.cls)
• Hardening: Refactored GetStats() to extract the structured https://claimauditai.com/fhir/extension/tier-results extension directly from the pended ClaimResponse resource.
• Accuracy: Replaced brittle substring parsing of the human-readable disposition text with deterministic JSON parsing of the flagged status for tier1, tier2, and tier3. This ensures the dashboard stats donut chart accurately counts flags regardless of formatting adjustments.

💬 Conversational Terminology Lookup

• Exposed FastMCP tools (lookup_cpt_code, lookup_icd_code, validate_clinical_edits) to the Chat Agent loop. When a user asks about a CPT/ICD code description in the UI Chat Assistant, the LLM issues a tool call to query the local terminology server, returning official definitions instead of generating hallucinations.

---

🚦 Verification Status

• Python Unit Tests: All 92 tests passed successfully, including the validation split, drift check, and backup rotation tests:

  $ pytest src/python/tests/
  ======================== 92 passed in 85.85s =========================
  

• Vite/React Unit Tests: Vitest suite completes successfully with 18 passed tests (100% success rate).
• End-to-End Workflow: Real-world integration tests successfully verify data clearing, seeding, retraining, audit hold escalation, director approvals, and localized graph analysis.

ClaimAuditAI — Release Notes v2.0.0 (Agentic Payment Integrity)

Welcome to the v2.0.0 Release of ClaimAuditAI. This release marks a major shift, transitioning the payment integrity engine from a sequential ReAct tool-calling loop into a type-safe, compiled Agentic Finite State Machine (FSM) powered by Pydantic Graph.

Additionally, this version incorporates advanced InterSystems IRIS for Health integrations, custom Model Context Protocol (MCP) terminology engines, and Explainable AI (XAI) evidence citation linkages.

---

🛠 Major Architectural Upgrades

1. Pydantic Graph FSM & Agentic Memory

We refactored the Python-based adjudication pipeline into a type-safe, compiled FSM (agent_graph.py):

• State-driven Adjudication (AuditState): Explicitly type-checked execution context tracking input metadata, tier-specific findings, citations, and LLM synthesis results.
• Structured Nodes: Implemented as independent classes inheriting from BaseNode[AuditState, None, str] with StepContext validation:
  • ClaimIngestionNode: Ingests and sanitizes metadata.
  • ClinicalAuditNode: Executes vector database clinical note alignment.
  • AnomalyAuditNode: Calculates autoencoder reconstruction loss.
  • NetworkAuditNode: Evaluates referral loops and address collisions on the provider graph.
  • LLMSynthesisNode: Aggregates scores and builds the markdown report.
• Deterministic Fast-Approval Bypass: Claims returning clean metrics across all three tiers bypass the LLM synthesis node entirely, transitioning directly to End to save API tokens and reduce latency.

2. Hybrid Agent Bridge Wrapper

• Class: ClaimAudit.AI.AgentWrapper
• Integration: Dynamically checks for the presence of native InterSystems %AI.Agent classes via class dictionary compilation. If present, it maps tools as %AI.Tool bindings under %AI.ToolSet to run natively within AI Hub. Otherwise, it falls back to execution via the Python-compiled Pydantic Graph FSM.

3. Unified Python Tool Registry

• File: [agent_tools.py](file:///Users/mck/Desktop/claimauditai/src/python/agent_tools.py)
• Implementation: Created a decorator-based tool registry (@registry.register) that dynamically generates OpenAPI/JSON schemas using function inspection. Exposes 8 medical and diagnostic tools:
  • lookup_cpt_code / lookup_icd_code (resolves medical codes).
  • validate_clinical_edits (verifies diagnosis-procedure combinations).
  • run_nlp_audit / run_anomaly_audit / run_graph_audit (live tier executions).
  • get_patient_history / get_provider_history (historical database lookups).

---

💡 Contest & InterSystems Enhancements

1. Terminology Model Context Protocol (MCP) Server

• Implementation: Developed [mcp_server.py](file:///Users/mck/Desktop/claimauditai/src/python/mcp_server.py) utilizing FastMCP to expose terminology resolution services.
• Functionality: Allows LLM models and the UI chat assistant to resolve medical codes dynamically and validate clinical relationships (CPT/ICD mapping).

2. FHIR SQL Builder Integration

• Projections: Schema definition file [fhirsql/projections.json](file:///Users/mck/Desktop/claimauditai/fhirsql/projections.json) enables standard projected tables for Claim and ClaimResponse resources.
• Helper Class: Compiled [SQLBuilderHelper.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/FHIR/SQLBuilderHelper.cls) to execute dynamic SQL statements over projections, eliminating manual JSON parsing queries.

3. Explainable AI (XAI) Evidence Citations

• Evidence Capturing: Modified the Python ML tiers to extract and bubble up absolute resource ID references (e.g. DocumentReference ID for clinical notes, Practitioner NPIs for address collisions).
• Database Mapping: Added the ClaimId column to the ClaimProjections database table, populated it during ingestion, and stored array citations in the FHIR tier-results extension on ClaimResponse.
• UI Presentation: Displays evidence citations as styled pill tags with hover effects underneath the tier results on the claim detail page.

---

🔒 Production Hardening & Bug Fixes

• Concurrency Safety: Replaced multi-threaded execution in tier_orchestrator.py with sequential processing to eliminate IRIS Embedded Python database context conflicts.
• Autoencoder Scaling: Scaled SpecialtyCode to [0,1] prior to evaluation, preventing categorical outliers from skewing continuous Z-score distances.
• JIT LLM Adjudication: Configured bulk seeding to bypass LLM generation, compiling detailed markdown reports on-demand (JIT) when an auditor actually views the claim detail.
• Ledger Linkage: Added claimResponseId to ledger records, converting Claim IDs in the Override Ledger into navigation links back to resolved read-only detail views.

---

🚦 Verification Status

• Python Unit Tests: All 90 pytest tests pass cleanly, including the new Pydantic Graph node transition and mock fallback suites.
• Vite/React Unit Tests: Vitest suite completes successfully with coverage reporting enabled.
• TypeScript & Lints: Clean compilations (tsc --noEmit exits with 0).
• End-to-End Workflow: Real-world integration tests successfully verify data clearing, seeding, retraining, audit escalation, and director decisions.

Here is the release description for ClaimAuditAI v1.0.2, formatted for immediate use in GitHub Releases, community announcements, or project documentation:

---

🚀 ClaimAuditAI v1.0.2-Enhancements Release

This release introduces three key architectural enhancements: FHIR SQL Builder Integration, a dedicated Medical Terminology MCP Server, and Explainable AI (XAI) Citations to link adjudication findings directly back to their source clinical and administrative evidence.

---

🌟 Key Highlights

1. ⚙️ FHIR SQL Builder Integration

Allows developers and administrators to bypass manual database projections in favor of platform-native relational maps for advanced analytics.

• fhirsql/projections.json: Pre-configured JSON projection configuration for Claim and ClaimResponse resources. This file can be imported directly into the InterSystems FHIR SQL Builder Management Portal.
• ClaimAudit.FHIR.SQLBuilderHelper: A custom ObjectScript query helper that executes dynamic SQL queries against projected schemas, extracting CPT codes, ICD diagnoses, amounts, and dates with built-in schema fallbacks.

🤖 2. Medical Terminology FastMCP Server

A Model Context Protocol (MCP) server built with Python's FastMCP framework, enabling cognitive agents to perform standard clinical translations and diagnosis-procedure justification checks.

• lookup_cpt_code(code): Resolves a 5-digit CPT code into its human-readable clinical procedure description.
• lookup_icd_code(code): Translates an ICD-10 diagnosis code into its standard clinical description.
• validate_codes(icd_code, cpt_code): Performs programmatic validation to ensure the diagnosis supports the billed procedure.

🔍 3. Explainable AI (XAI) Citations

Links threat signals back to specific, traceable FHIR resource identifiers in the adjudication trail, giving auditors exact references to support hold decisions.

• Tier 1 (NLP): Captures the DocumentReference/{id} identifier of the progress note matching the billed CPT code.
• Tier 2 (ML): Citations link directly back to the audited Claim/{id} as evidence of statistical outlier loss.
• Tier 3 (Graph): Returns conflicting Practitioner/{npi} and Claim/{id} references for geodetic/temporal leaps or address collisions.
• UI Integration: Displays evidence citations as clean, monospace pill tags featuring color-coded indicator dots matching the active tier and subtle scale transitions on hover.

---

🛠️ Technical Changelog

🗄️ Database & ObjectScript Layer

• Updated [Engine.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/AI/Engine.cls) table schema definition to add ClaimId to the ClaimProjections table with a safe ALTER TABLE schema update fallback.
• Updated [Router.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/REST/Router.cls) and [Engine.cls](file:///Users/mck/Desktop/claimauditai/src/cls/ClaimAudit/AI/Engine.cls) to parse python citations lists and persist them into the dynamic tier-results extension array.

🐍 Python Adjudication Pipelines

• Added mcp>=1.27.2 to [requirements.txt](file:///Users/mck/Desktop/claimauditai/requirements.txt).
• Updated [nlp_auditor.py](file:///Users/mck/Desktop/claimauditai/src/python/nlp_auditor.py) to select and return note references DocumentReferenceId in native vector search.
• Updated [graph_analyzer.py](file:///Users/mck/Desktop/claimauditai/src/python/graph_analyzer.py) to read ClaimId from projected edges and return conflicting NPIs and claims for address collisions.
• Updated [tier_orchestrator.py](file:///Users/mck/Desktop/claimauditai/src/python/tier_orchestrator.py) to pass claim_id parameters and forward citations downstream.

💻 React UI Dashboard

• Added citations type interfaces in [claim.ts](file:///Users/mck/Desktop/claimauditai/ui/src/types/claim.ts).
• Styled and integrated citations rendering in [TierPanel.tsx](file:///Users/mck/Desktop/claimauditai/ui/src/components/claims/TierPanel.tsx) with responsive padding and micro-interactions.

📚 Documentation & Wiki

• Added the [Terminology MCP Server](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/07-api/Terminology%20MCP%20Server.md) wiki page.
• Updated the [FHIR SQL Builder Projections](file:///Users/mck/Desktop/claimauditai/claimauditai-wiki/04-fhir/FHIR%20SQL%20Builder%20Projections.md) wiki page and the MOC home index.

---

🚦 Verification Status

• Python Tests: 82 passed unit tests (including the new test_mcp_server.py suite).
• UI Code Quality: TypeScript check compiled cleanly with zero errors (npx tsc --noEmit exited with 0).
• Workflow integration: Successfully executed real_world_e2e_tests.py against the running docker containers to verify database seeding, autoencoder training, single-claim details fetching, and correct REST response mapping.

• missing module.xml update ...