Skip to content

Features

Vigilyx is built as an engineering-oriented email security platform. The feature set is organized around deployment flexibility, explainable detection, operational response, and long-term maintainability.

Deployment shapes

  • Mirror mode: passively capture SMTP, POP3, IMAP, and webmail traffic without changing the existing mail path
  • Inline MTA proxy mode: accept, quarantine, relay, or reject mail before delivery
  • Shared detection stack: the same parsing, detection, and verdict logic can run behind both topologies
  • Optional AI sidecar: NLP enrichment is available, but the Rust core still works when AI is disabled

Parsing and session analysis

  • SMTP and MIME parsing
  • Passive mail-session reconstruction from captured traffic
  • Attachment extraction and metadata normalization
  • Header, HTML, link, and attachment context building
  • Reusable parsing logic shared by the sniffer and the MTA proxy

Detection pipeline

  • Multi-module threat analysis pipeline
  • Header, content, semantic, link, reputation, MIME, HTML, and anomaly modules
  • Attachment heuristics and hash-based detection
  • YARA scanning for messages and attachments
  • DLP checks for sensitive data and data-security workflows
  • Verdict fusion with explainable, reviewable scoring

Response and enforcement

  • Risk levels from safe to critical
  • Inline accept / quarantine / reject behavior in MTA mode
  • Quarantine management and release workflows
  • Alerting, webhooks, and automated response hooks
  • Audit-oriented session and verdict history for analysts

Operations and product surface

  • Real-time React dashboard
  • REST API and WebSocket updates
  • Docker Compose deployment for mirror and MTA topologies
  • Remote-first deploy.sh workflow for fast developer packaging and production releases
  • Bilingual public docs and GitHub Pages project site

Design priorities

  • Explainable verdicts instead of black-box-only scoring
  • Clear separation between control plane and traffic-processing data planes
  • Practical deployment on CPU-only infrastructure
  • Rules, IOCs, and fusion behavior that can be iterated over time

Released under AGPL-3.0-only.