Linux / macOS

Single command

Install and run in one step with curl. The CLI collects a sanitized diagnostic snapshot (OS, packages, ports, firewall, Docker, databases, TLS certs) and sends it for AI analysis. No passwords, keys, or file contents are ever touched.

Copycurl -fsSL https://taloscloud.net/scan | bash

After install you can also just run:

Copytalosscan

Options

Copytalosscan --yes      # skip confirmation prompt
talosscan --dry-run  # show what would be sent
talosscan --install  # only install, don't scan
talosscan --help     # show full usage

Windows (PowerShell)

Single command

Run in an elevated (administrator) PowerShell terminal for full system access. The scanner collects the same diagnostics — OS, network, firewall, updates, listening ports, and more.

Copyirm https://taloscloud.net/scan.ps1 | iex
Run in an elevated PowerShell (right-click → Run as Administrator) for the most complete results.

AI Chat

Interactive

Every report comes with an AI assistant you can talk to. Ask about findings, enable/disable suggested fixes, or ask it to run live diagnostic commands on your machine.

After getting your report link, just run:

Copycurl -fsSL https://taloscloud.net/aichat/<REPORT_ID> | bash

The AI can run live commands via [[RUN: ...]] to check disk, processes, or anything else — you see both the command and its output.

What gets collected

Privacy-first

The snapshot is fully sanitized — no secrets, keys, passwords, tokens, or file contents are ever included. Here's exactly what the scanner looks at:

System
OS, kernel, uptime, CPU, memory, disk
Network
Interfaces, IPs, DNS, open ports with service names
Security
Firewall rules, SSH config, Fail2Ban, pending updates
Docker
Containers, privileged mode, socket mounts
Databases
PostgreSQL, Redis, MySQL — localhost vs public
TLS
Local HTTPS cert expiry dates
Packages
Version audit for openssl, nginx, docker, php, python
Exposure
Internet-facing services and attack surface rating

After analysis you get a private report link with a posture score (0–100), categorized findings, suggested fixes, and an AI you can talk to interactively.