| Tool | Purpose | How to Use |
|---|---|---|
| 37-data-model | Single source of truth for all project entities | GET http://localhost:8010/model/projects/27-devbench |
| 29-foundry | Agentic capabilities (search, RAG, eval, observability) | C:\eva-foundry\eva-foundation\29-foundry |
| 48-eva-veritas | Trust score and coverage audit | MCP tool: audit_repo / get_trust_score |
| 07-foundation-layer | Copilot instructions primer + governance templates | MCP tool: apply_primer / audit_project |
Agent rule: Query the data model API before reading source files.
Invoke-RestMethod "http://localhost:8010/model/agent-guide" # complete protocol
Invoke-RestMethod "http://localhost:8010/model/agent-summary" # all layer counts
AI-Assisted Software Engineering Platform for Legacy Modernization
EVA DevBench is a production-quality web application designed for government engineering teams modernizing legacy applications. It orchestrates AI-assisted work packages (reverse engineering, scaffolding, refactoring, testing, documentation) and produces auditable artifacts through secure, traceable runs.
Support developers reducing technical debt in application libraries based on 60+ years of legacy technology:
- COBOL mainframe applications
- Client-server architectures
- Three-tier legacy systems
- Untested or undocumented codebases
- Project Management: Organize modernization work with metadata tracking
- Context Bundle Builder: Securely package code with cryptographic sealing
- Work Package Catalog: Pre-configured AI workflows for common tasks
- Live Run Console: Real-time SSE event streaming with agent timelines
- Enhanced Diff Viewer: Syntax highlighting for 15+ languages, split/unified views
- Review Mode: Risk indicators (Low/Medium/High) on every file
- Agent Rationale: See why AI made each change with assumptions and confidence
- File Provenance: Complete audit trail (agent, model, context, run ID, timestamp)
- Semantic Grouping: Organize files by architectural layer (Controllers, Services, Tests)
- Review Status Markers: Mark files as OK/Needs SME/Needs Re-run with persistence
- "Explain This Hunk": AI-powered contextual explanations for code changes
- Protected Zone Tracking: View what files AI explicitly did NOT touch
- Word-Level Diffs: Character-by-character change highlighting
- Full Audit Trails: Trace IDs, timestamps, token counts, costs
- AICOE-Grade Provenance: Every line auditable to source
- Evidence Bundles: Exportable JSON reports for compliance
- Bilingual Support: EN/FR toggle with persistent preference
- WCAG 2.1 AA: Full keyboard navigation and accessibility
npm installnpm run devAccess at http://localhost:5173
- Navigate to Settings
- Toggle Demo Mode ON
- Explore with realistic mock data
See QUICKSTART.md for detailed workflows.
- QUICKSTART.md: 5-minute demo setup and core workflows
- EXPORT.md: Complete technical documentation and architecture
- DEPLOYMENT.md: Full Azure deployment guide with all options
- AZURE_DEPLOYMENT.md: GitHub Actions deployment step-by-step
- QUICKSTART_DEPLOY.md: Deploy in 5 minutes quick reference
- PRD.md: Product requirements and design decisions
- Framework: React 19.2.0 + TypeScript 5.7.3
- Build Tool: Vite 7.2.6
- Styling: Tailwind CSS 4.1.17 + Shadcn/ui v4
- Icons: Phosphor Icons React
- State: React hooks + useKV (persistent)
- HTTP: Fetch API with APIM integration
- Real-time: Server-Sent Events (SSE)
- Colors: OKLCH color space for perceptual uniformity
- Typography: Space Grotesk (headings), Inter (body), JetBrains Mono (code)
- Accessibility: WCAG 2.1 AA compliant
- Animations: Purposeful, subtle motion with Framer Motion
- No client-side secrets
- All API calls through Azure APIM gateway
- Bearer token authentication (Entra ID ready)
- Cryptographic bundle sealing
- Protected zone enforcement
Version: 0.1.0 (MVP)
Iterations: 20 completed
Status: Production-ready for Azure EsDAICoESub sandbox
Deployment Target: Azure Static Web Apps or App Service
- Reverse engineer mainframe business logic
- Generate modern .NET/Java equivalents
- Preserve exact calculation behavior
- Create comprehensive test suites
- Document three-tier architecture patterns
- Generate modern API layers
- Refactor data access to ORMs
- Create migration documentation
- Auto-generate tests for untested code
- Refactor legacy modules incrementally
- Document undocumented systems
- Modernize UI layers
This application is designed for secure government environments. Before contributing:
- Review SECURITY.md
- Ensure WCAG 2.1 AA compliance
- Follow TypeScript best practices
- Add translations for EN/FR
- Update PRD.md with feature changes
- Technical Documentation: See EXPORT.md
- Deployment Issues: See DEPLOYMENT.md
- Feature Questions: See PRD.md
- Training: See QUICKSTART.md
Spark Template files and resources from GitHub are licensed under the MIT license, Copyright GitHub, Inc.
Built for: EsDAICoE Sandbox
Maintained by: EVA DevBench Team
Export Date: 2024-01