How ejajaka/vector works
Vector diagnoses natural-language cloud-infrastructure prompts before code generation: its rule-based engine identifies resources and controls, reports omissions and risky statements, and can produce security clauses for an improved prompt. The CLI also exposes verification, compliance, cloud-specific analysis, and optional semantic, coreference, and LLM reasoning modes. The README documents extension, web, web-app, Expo, Capacitor, and native Android clients. Client internals are only partly sampled, so their detailed runtime links are not inferred beyond documented engine use.
Main entry points: analyzer.js Prompt analyzer · vector-cli.js CLI commands · cloud-analysis.js Cloud analysis · App.js Expo mobile app
Components 23
Analysis engine
- Prompt analyzer
engine/src/analyzer.js - Security taxonomy
engine/src/taxonomy.js - Cloud analysis
engine/src/cloud-analysis.js - Cloud controls
engine/src/taxonomy-clouds.js - Prompt improvement
engine/src/analyzer.js - Infrastructure checks
engine/src/tfcheck.js - Compliance checks
engine/src/compliance.js
Analysis extensions
- Deep NLP
engine/src/deep-nlp.js - Coreference linking
engine/src/coref-link.js - Semantic matching
engine/src/semantic-embed.js - LLM reasoning
engine/src/llm-reasoning.js - Optional deep scan
engine/src/llm.js
User interfaces
- Browser extension
ui/extension/popup.js
Mobile clients
- Expo mobile app
ui/mobile-expo/App.js - Mobile history
ui/mobile-expo/src/storage.js - Capacitor client
ui/mobile-native/www/app.js - Native Android
Web clients
- Static web demo
ui/web/app.js - React web app
ui/web-app/src/App.tsx - Web history
ui/web-app/src/services/historyStorage.ts
Other components
- Prompt author
- CLI commands
engine/cli/vector-cli.js - Configured AI service
Connections 30
- Prompt author → CLI commands: enters prompt
README.md - CLI commands → Prompt analyzer: analyzes
engine/cli/vector-cli.js - CLI commands → Cloud analysis: runs cloud scan
engine/cli/vector-cli.js - CLI commands → Infrastructure checks: verifies output
engine/cli/vector-cli.js - CLI commands → Compliance checks: checks standards
engine/cli/vector-cli.js - CLI commands → LLM reasoning: requests reasoning
engine/cli/vector-cli.js - CLI commands → Optional deep scan: requests deep scan
engine/cli/vector-cli.js - Prompt analyzer → Security taxonomy: uses controls
engine/src/analyzer.js - Prompt analyzer → Semantic matching: uses semantic pass
engine/src/analyzer.js - Prompt analyzer → Prompt improvement: builds clauses
README.md - Cloud analysis → Prompt analyzer: uses analysis
engine/src/cloud-analysis.js - Cloud analysis → Cloud controls: uses controls
engine/src/cloud-analysis.js - Coreference linking → Deep NLP: resolves references
engine/src/coref-link.js - Coreference linking → Prompt analyzer: reanalyzes prompt
engine/src/coref-link.js - LLM reasoning → Configured AI service: calls configured model
engine/src/llm-reasoning.js - Optional deep scan → Configured AI service: calls hosted model
engine/src/llm.js - Prompt author → Browser extension: uses
README.md - Prompt author → Static web demo: uses
README.md - Prompt author → React web app: uses
README.md - Prompt author → Expo mobile app: uses
README.md - Prompt author → Capacitor client: uses
README.md - Prompt author → Native Android: uses
README.md - Browser extension → Prompt analyzer: uses shared engine
README.md - Static web demo → Prompt analyzer: uses shared engine
README.md - React web app → Prompt analyzer: uses shared engine
README.md - Expo mobile app → Prompt analyzer: uses shared engine
README.md - Capacitor client → Prompt analyzer: wraps web client
README.md - Native Android → Prompt analyzer: uses shared engine
README.md - React web app → Web history: stores analyses
- Expo mobile app → Mobile history: stores analyses
ui/mobile-expo/App.js