Andris Enins
91cd18aec6
Backend (Spring Boot 3.2 / Java 21 / PostgreSQL): - JWT auth with BCrypt password hashing - User profile + Mifflin-St Jeor BMR calculator - Food search + barcode via OpenFoodFacts API with local cache - Meal CRUD with user data isolation and ownership checks - AI photo analysis (OpenAI Vision) with confidence intervals - AI correction feedback loop for personalisation - Flyway DB migrations + RFC-7807 error responses Mobile (React Native / TypeScript): - Full navigation stack (Auth → Tabs → Home stack) - Design tokens (WCAG 2.2 AA colours, 8px grid, 48px touch targets) - 10 screens: Login, Register, Home, Search, Camera, AI Result, Edit Meal, Daily Details, History, Profile - Confidence-aware calorie display (kcal ± range) - Repeat last meal shortcut + macro tracking Docs: - docs/PLAN-AND-REQUIREMENTS.md - docs/traceability.csv (35 requirements, all Implemented)
6.2 KiB
6.2 KiB
Virsaitis Requirements - Index
Project: Virsaitis Three-Layer AI Governance System Version: 3.0.0 Status: In Development Owner: Toms Eisaks Updated: 2026-04-21
📋 Requirements Overview
Total requirements: 83 across 7 categories
| Category | Count | REQ-ID Range | Status |
|---|---|---|---|
| Governance Core | 12 | REQ-GOV-001 to 012 | Draft |
| Security Controls | 16 | REQ-SEC-001 to 016 | Draft |
| MCP Server | 11 | REQ-MCP-001 to 011 | Tested (277 tests, 100% functions) |
| VS Code Extension | 21 | REQ-EXT-001 to 021 | Tested (136 tests, 83% statements) |
| Agent | 8 | REQ-AGT-001 to 008 | Implemented |
| Skills | 5 | REQ-SKL-001 to 005 | Draft |
| Testing & QA | 10 | REQ-TEST-001 to 010 | Draft |
🎯 Project Mission
Create three-layer AI governance system that achieves 95%+ compliance for AI-assisted software development through:
Layer 1: Agent (Behavioral Guidance)
- Atomic markdown instruction design
- Self-regulation through clear rules
- Consequence-aware decision making
Layer 2: MCP Server (Pre-execution Validation)
- TypeScript governance enforcement engine
- File operation validation
- Secret scanning and input validation
Layer 3: VS Code Extension (User Action Interception)
- Real-time file save interception
- Visual governance indicators
- Override workflow management
Layer 4: Skills (Domain-Specific Rules)
- Native VS Code Agent Skills
- Progressive disclosure (3 levels)
- Consequence documentation per TIER
📚 Requirements Documents
Core Requirements
- functional-spec.md - Functional requirements for all components
- nonfunctional-spec.md - Performance, scalability, usability
- security-controls.md - Security requirements and controls
- testing-requirements.md - Test coverage and quality gates
Supporting Documents
- glossary.md - Terminology and definitions
- assumptions.md - Project assumptions log
- risk-register.md - Identified risks and mitigations
- traceability.csv - REQ-ID to Implementation mapping
🔑 Critical MUST Requirements (TIER-0)
These requirements are non-negotiable and block production deployment if not met:
- REQ-GOV-001: Protected file modification enforcement
- REQ-GOV-002: Atomic sentence structure in Agent.md
- REQ-SEC-001: Secret detection 100% coverage
- REQ-SEC-002: Credential rotation within 1 hour
- REQ-MCP-003: File operation validation engine
- REQ-EXT-002: File save interception for protected files
- REQ-TEST-001: Security test coverage 100%
📊 Requirements by Priority
TIER-0 (Safety-Critical) - 12 requirements
Must be 100% implemented and verified. No exceptions.
TIER-1 (Code-Breaking) - 28 requirements
Must be ≥95% implemented. Minimal compromise allowed with approval.
TIER-2 (Quality Standards) - 21 requirements
Should be ≥80% implemented. Acceptable tradeoffs with justification.
TIER-3 (Enhancements) - 10 requirements
Best effort implementation. Negotiable based on resources.
🚀 Implementation Phases
Phase 1: Foundation (Complete)
- Agent.md with atomic sentences
- 11 governance modules
- Core requirement documents
- Traceability framework
Phase 2: MCP Server (Complete)
- TypeScript MCP server implementation (14 source files, 2,799 LOC)
- Governance validation engine (8 tools)
- Secret scanning (Shannon entropy + regex patterns)
- Input validation and rate limiting
- Test suite: 277 tests, 100% function coverage
Phase 3: VS Code Extension (Complete)
- File save interception (readonlyInclude + post-save revert)
- Visual governance indicators (shield badges, status bar)
- MCP client (stdio transport, lifecycle management)
- Override workflow and audit trail
- Framework install/detect/update/validate commands
- First-run setup wizard
- Test suite: 136 tests, 83% statement coverage
- VSIX packaged: 688 KB, distributed to virsaitis-distribution/
Phase 4: Skills Development
- 6 core skills with Consequences sections
- Progressive disclosure implementation
- Skills validation
- Integration testing
Phase 5: Portable Distribution
- Build automation
- Installation scripts (Windows/Linux/Mac)
- Documentation
- Distribution packaging
- User acceptance testing
📈 Success Metrics
| Metric | Target | Current | Measurement |
|---|---|---|---|
| Governance Compliance | ≥95% | — | TIER-0 violations per month |
| MCP Test Coverage | ≥70% | 100% functions | virsaitis-mcp: 277 tests |
| Extension Test Coverage | ≥80% | 83% statements | virsaitis-extension: 136 tests |
| Security Test Coverage | 100% | 100% | Secret scanning + TIER-0 enforcement |
| Requirement Coverage | 100% MUST | 100% EXT, 100% MCP | 112/112 EXT ACs, all MCP tested |
| False Positive Rate | <5% | — | Incorrect blocks |
| Response Time | <10s | <10s | MCP timeout default |
| VSIX Size | <10MB | 688 KB | virsaitis-3.0.0.vsix |
🔗 Related Documentation
- Architecture: ../virsaitis-documentation/5-COMPONENT-ARCHITECTURE.md
- Agent Standards: ../.github/copilot-modules/agent-standards.md
- MCP Standards: ../.github/copilot-modules/mcp-standards.md
- Extension Standards: ../.github/copilot-modules/extension-standards.md
📝 Change History
| Date | Version | Author | Changes |
|---|---|---|---|
| 2026-04-21 | 3.0.0 | Toms Eisaks | EXT status → Tested (136 tests, 112/112 ACs), phases 1-3 complete, metrics updated |
| 2026-04-20 | 3.0.0 | Toms Eisaks | MCP status → Tested (277 tests), Agent → Implemented, traceability populated |
| 2026-02-17 | 2.0.0 | Toms Eisaks | Initial requirements structure for Virsaitis v2.0 |
Virsaitis Requirements Index v3.0.0 Three-layer AI governance system — Phases 1–3 complete