Open research.
Open code.

Open source host scanner, developed and maintained by S4E. A collector reads the software installed on a Linux or Windows server, a resolver matches it against vendor advisories and explains every finding. Inside depth in S4E builds on it.

The half of Openness that runs on the host. S4E delivers it over SSH or WinRM, or you run it locally; it reads installed packages, OS release and running services into a signed evidence file. Read only, no agent left behind. Developed and maintained by S4E.

The half of Openness that explains. Matches collector evidence against vendor advisories (Ubuntu USN and OVAL, Alpine secdb, OSV, Microsoft MSRC) and produces findings with the installed version, the fixed version and the advisory. Developed and maintained by S4E.

MCP server that brings the full S4E platform into your AI assistant. Connect Cursor, Claude Desktop, Claude Code, VS Code, or Windsurf and manage assets, launch Full/Light/Crawler scans, read findings, tune crawling, and review posture, all in natural language over HTTP. Per-user API token auth; no credentials stored on the server.

Real-time security alert client. Monitors your assets continuously and fires notifications the instant a new vulnerability appears, a port opens, or a certificate expires.

Pre-built scan sequences for compliance, pentest, and reconnaissance. PCI-DSS, ISO 27001, full attack surface mapping. Select a goal, hit run, S4E does the rest.

A security layer between AI coding agents and your shell. Every command Claude Code, Cursor, or Codex suggests passes through ShellGuard first: logged, policy-checked, and blocked or allowed. Single Go binary, sub-millisecond overhead.

Native macOS menu bar app for ShellGuard. One-click hook setup for every AI tool, live activity log, mode switching, and a policy editor, all from the menu bar. The security logic lives in the CLI; this is the GUI layer on top.