Skip to main content

JavaScript and TypeScript

Vulkro does full source analysis for JavaScript and TypeScript. It covers the Node.js server frameworks and the client-side XSS / DOM-sink surface.

Both languages are Free, like every language, with no upgrade needed for the deep analysis and no account needed to scan. See Accounts, Free and Pro.

Project detection​

Vulkro spots the project from package.json and reads installed versions from package-lock.json. TypeScript is analysed directly. There is no build step or type-check to run first.

vulkro scan .
vulkro discover .            # endpoint map only

Frameworks​

Vulkro extracts endpoints from:

  • Express
  • Fastify
  • NestJS
  • Next.js (App Router and Pages Router)
  • Koa
  • Hono
  • Hapi
  • Elysia
  • AdonisJS
  • tRPC

It also covers:

  • GraphQL (Apollo, graphql-yoga)
  • gRPC
  • WebSocket (ws, Socket.IO)
  • React and Angular client surfaces (for XSS and DOM-sink analysis)

Router composition is followed so the recorded path matches what the server serves (for example Express Router mounts and Next.js route segments).

Access model​

Every endpoint is tagged with an AuthTier (Anonymous / User / Admin / Service / Machine / Unknown), its scopes, and its tenant-scoping. The BOLA / IDOR and broken-authentication rules read this metadata, so a finding already handled by a visible auth gate is suppressed. Java, Python, and Go share the same endpoint model, and all of them feed the OWASP API Top 10 checks.

Taint analysis​

Vulkro follows request data through function calls to dangerous sinks.

Sources include req.body, req.query, req.params, req.headers, the equivalents in Next.js route handlers, and route-handler parameters.

Sinks include:

  • SQL query builders (knex.raw, sequelize.query, db.query, pool.query)
  • child_process.exec
  • eval / Function(...)
  • fetch(...) (SSRF)
  • redirect(...) (open redirect)
  • template-as-string XSS
  • __proto__ writes (prototype pollution)

Each finding shows the source-to-sink hop chain (SARIF codeFlows).

Loop bodies are walked once, so loop-carried taint (a value that re-enters a variable on a later iteration) is not tracked. See Taint analysis for the full source / sink / sanitiser model and its limits.

Dependencies​

Vulkro reads npm dependencies from package.json (resolved via package-lock.json) and matches them against the local CVE bundle.

For a curated set of high-traffic packages, CVE findings are tagged [reachable] or [unreachable] based on whether the vulnerable code is actually called. See Dependencies and CVEs and Reachability.

The AI-generated-code checks (slopsquatted package names, RAG-index safety, CI/CD pipeline injection) apply to JS/TS. See New detectors.

Useful commands​

vulkro scan .
vulkro discover . --format json
vulkro dataflow .
vulkro explain JS-TAINT-SQL-001
Last updated on