Process — Sovereign Forge

// Phase 01

Discovery

The Discovery phase tears apart legacy application packages to build a machine-readable system manifest. Every installer format, every binary, every database schema, every network call — captured, catalogued, and documented automatically.

The output is a comprehensive technical blueprint of what the application actually does at the system level, independent of any documentation that may or may not exist.

→Installer Extraction — Automated unpacking of MSI, NSIS, Inno Setup, and InstallShield packages. Every file, registry key, and service definition extracted.
→Binary Analysis — Ghidra headless mode performs function-level decompilation. Call graphs, string references, import tables, and API usage patterns identified.
→Database Schema Recovery — SQL Server, Access, and FoxPro databases reverse-engineered. Table relationships, stored procedures, triggers, and data types mapped.
→Registry & Filesystem Diff — Before/after snapshots capture every system modification the application makes during installation and runtime.
→Network Protocol Capture — All external communications recorded. HTTP/HTTPS endpoints, database connections, SMTP, and custom protocols identified.
→Deployment Documentation — Auto-generated installation guide based on observed behavior, not vendor claims.

📦

Installers Extracted

MSI, NSIS, Inno Setup, InstallShield

🔬

Binaries Analyzed

Ghidra headless decompilation

🗄️

Schemas Recovered

SQL Server, Access, FoxPro

🌐

Protocols Captured

HTTP, SQL, SMTP, custom

Output → Machine-readable system manifest

// Phase 02

Knowledge Capture

The most critical business logic often lives in people's heads, not in code. The Knowledge Capture phase extracts institutional knowledge through structured interviews, screen recordings, and AI-assisted workflow analysis.

This is where the 30-40% that can't be automated gets addressed — the domain semantics, the "we always do it this way" workflows, and the undocumented business rules that no decompiler will ever find.

→Guided Interviews — Structured questionnaires for administrators and end users, recorded and transcribed via Whisper. Role-specific question trees adapt based on responses.
→Screen Recording — User sessions captured with AI annotation via Azure AI Vision. Every click, every navigation path, every data entry pattern documented.
→Workflow Mapping — Role-based process flows derived from observed behavior. End-to-end business processes mapped from initiation to completion.
→Business Rule Extraction — The logic that lives in people's heads: exception handling, edge cases, seasonal variations, approval chains, and informal procedures.

Output → Functional specification + process maps

// Phase 03

Analysis & Decision

Phase 3 correlates everything discovered in Phases 1 and 2 into an actionable BI dashboard. Complexity scoring, risk heat maps, workflow coverage analysis, and integration surface area — all visualized for decision-makers.

The platform recommends a modernization target but the human makes the call. Every recommendation comes with cost, effort, and risk modeling.

→Complexity Scoring — Weighted analysis of code complexity, integration density, data sensitivity, and user workflow coverage.
→Risk Heat Maps — Visual identification of high-risk components: undocumented integrations, complex business logic, compliance-sensitive data flows.
→Compliance Mapping — Gap analysis against CMMC, HIPAA, PCI-DSS, and NIST 800-171. Requirements mapped to specific application components.
→Target Recommendation — Cost/effort/risk model for each modernization path: SaaS replacement, internal web app, fat client 1:1, or hybrid.

☁️

SaaS Replacement

Commodity functionality, low customization

🌐

Internal Web App

Org-specific logic, stays on-prem

🖥️

Fat Client 1:1

Preserve user workflows exactly

⚡

Hybrid

Web + containers + preserved components

Output → Target recommendation with gap report

// Phase 04

Generation & Validation

The final phase produces deployable artifacts. AI agents generate microservice scaffolding based on the discovered specifications. Test suites are derived from captured workflows. Everything is validated against the original application running in a sandboxed environment.

The output is not a black box — it's source code, container images, API specifications, and test suites that the organization owns and can maintain.

→Code Scaffolding — Semantic Kernel and AutoGen agents generate microservice code from discovered specifications and captured business rules.
→OpenAPI Generation — RESTful API specifications generated from discovered endpoints and data contracts.
→Test Suite Generation — Automated test cases derived from captured user workflows. End-to-end, integration, and regression coverage.
→Container Packaging — Docker/OCI images, Helm charts, and AKS manifests. Ready for deployment on AKS Hybrid.
→Regression Validation — Sandboxed comparison of modernized output against original application behavior. Side-by-side functional verification.

Output → Deployable containers + comprehensive gap report

Four phases. Full automation.Human validation at every gate.