Back to portfolio Completed demo

Sentinel Caseworks

A secure local workspace for sensitive investigations—connecting authority, people, evidence, chronology, tasks, reports, and integrity controls without making the cloud the default.

Project summary

Case discipline, evidence integrity, and local control in one operating record.

Industry
Private investigation / sensitive casework
Architecture
PySide6 desktop application -> SQLAlchemy -> encrypted local SQLite workspace
Primary tools
Python, PySide6, SQLAlchemy, SQLite, AES-256-GCM, Argon2id, ReportLab
Deployment model
Offline-capable Windows one-folder application
Purpose
Case intake, authority review, entities, encrypted evidence, timelines, tasks, reporting, audit, backup, and optional read-only local AI
System architecture

Built for traceability before convenience.

Sentinel keeps the core workspace local and makes authority, source attribution, custody, and immutable activity visible inside the daily workflow.

Local Desktop App

Windows workspace designed to operate without an Internet dependency.

Encrypted Records

Protected sensitive text and evidence with authenticated encryption.

Integrity Chains

Append-only custody, notes, reviews, and audit events with verification.

Local AI Option

Read-only case assistance through an already-installed local model.

01

Authority & Compliance

Record client authority, permissible purpose, jurisdiction, restrictions, license context, review decisions, and next-review dates.

02

People & Entities

Connect people, organizations, locations, and vehicles to confidence levels and source attribution.

03

Encrypted Evidence

Import protected files, calculate SHA-256 values, maintain custody events, and verify evidence integrity.

04

Timeline & Notes

Build sourced chronology while keeping contemporaneous notes append-only and attributable.

05

Relationships & Tasks

Map connected entities, organize assignments, and keep open work visible inside the active case.

06

Reports & Audit

Generate professional PDF summaries, inspect integrity results, and maintain encrypted backup and restore workflows.

Video demonstration

Follow one fictional case through the complete workflow.

The walkthrough covers authority review, sourced entities, protected evidence, chronology, relationships, tasks, reporting, audit integrity, and the optional local assistant.

Security posture

Local-first does not mean risk-free. It means the controls are explicit.

The demo materially reduces exposure through Windows-protected keys, authenticated encryption, strong password hashing, restricted local permissions, audit chains, encrypted backups, and offline deployment guidance. Endpoint compromise, administrator access, insiders, and physical access still require operational controls.

Evidence
AES-256-GCM
Passwords
Argon2id
Integrity
Hash chains
AI mode
Local / read-only
Case-study facts

A completed demonstration with clear deployment boundaries.

Status
Working Windows desktop demonstration with packaging and automated tests.
Data
One fully fictional Oklahoma workers' compensation review; no real subject or client information.
Customization
Product, organization, brand, tagline, application identity, security policy, and AI choice are configurable before delivery.
AI boundary
Optional local read-only assistant; no automatic downloads and no mutation tools.
Legal boundary
Workflow and recordkeeping support only—not legal advice, a license to investigate, or a guarantee of compliance or outcome.
Distribution
One-folder Windows build; the complete folder is required, not the executable alone.
Technology

Technologies used in Sentinel Caseworks.

  • Python
  • PySide6
  • SQLAlchemy
  • SQLite
  • AES-256-GCM
  • Argon2id
  • Windows DPAPI
  • ReportLab
  • Ollama
  • PyInstaller
  • pytest

Need a serious local application for sensitive work?

OCBS builds focused Windows systems around the workflow, data boundaries, and handoff needs of small local organizations.