Security model
What the InfraFence agent accesses on your server and why, what it sends to the dashboard, how tokens are stored, how releases are signed, the AI features and the BSL 1.1 license.
Last updated:
Running a security agent means giving it root on your servers. This page says exactly what it does with that.
Why root
The agent manages the firewall (iptables and ipset), reads the system's logs and network state, and applies the fixes you ask for. Those need root. It runs as one static binary with no dependencies, as a service you can stop at any time.
What it reads
| What | Why |
|---|---|
The SSH authentication log (systemd journal, /var/log/auth.log or /var/log/secure) |
Brute force, session risk |
| Web server access logs, ModSecurity audit log | Web firewall, bots |
| Mail, database and FTP logs | Brute force on those services |
DNS packets (UDP port 53, read-only) and connection tables in /proc |
Outbound and DNS threats, port scans |
Website files, including .env and configuration files |
Malware scan, framework checks |
WordPress posts, options and users (with the credentials in wp-config.php) |
WordPress database scan |
| Running processes | Crypto miners, reverse shells |
| Watched system files (accounts, sudoers, SSH keys, cron, systemd units) | File integrity |
| Installed packages and versions | Vulnerabilities (CVE) |
| The Docker socket, read-only, when Docker is present | Container logs and ports |
What it changes
- Its own firewall chains and ipsets. Your rules are never modified, and it never adds an
ACCEPT. - Only when you turn Allow web server configuration changes on: its ModSecurity rules and user-agent blocking in nginx or Apache.
- Only when you, AutoMode or the emergency mode ask: hardening fixes in files of its own, closed ports, outbound blocks, virtual patches, security updates, quarantined files.
Every change to the server is verified afterwards and undone if SSH or a site stops answering. Uninstall removes all of it.
What it sends
Everything goes to the dashboard over HTTPS:
- a heartbeat every 60 seconds: agent version, address, system metrics (CPU, memory, disk, load, network), listening services and ports, the web server and the domains it serves, firewall counters;
- security events: attacker address, type, time and the details needed to understand them, for a web attack the single log line that triggered it, for malware the file path and the matched text;
- the list of installed packages, for the vulnerability check;
- the SHA-256 of scanned PHP files, to look them up among known malware files.
Full logs and files never leave the server. No telemetry, nothing shared with third parties, except for the AI features below.
Tokens
- The install token is single-use, valid one hour, and is not kept on the server after registration.
- The agent's own token is stored in
/etc/infrafence/config.json, readable only by root. The dashboard keeps only a hash of it, so a copy of the dashboard's database does not contain usable agent tokens. Cluster keys for Kubernetes are also stored hashed and shown only once. - Changes from the dashboard arrive over a real-time channel that only tells the agent to synchronize: the agent then fetches its configuration over HTTPS with its own token. A message on that channel is never trusted as an instruction.
Signed releases
Every release is built by GitHub Actions, signed with Cosign and published with build provenance. The agent installs an update only if both its checksum and its signature verify against the public key built into it, and only from the official GitHub releases. See Verify a release.
AI features
The AI features run in the dashboard, never in the agent. When you ask for an AI analysis, use the assistant or draft a report, and while AutoMode or the emergency mode is on, the events needed are sent to Anthropic to produce the answer. The dashboard says so every time before you send. The instant analysis, available on every plan, sends nothing anywhere.
When the dashboard is unreachable
Protection stays on. See When the dashboard is unreachable.
License
The agent is source-available under the Business Source License 1.1 (licensor: ComCart S.r.l.). Production use is free on any number of servers, including your customers' servers, as long as you do not use it to build a competing product. Four years after each release, that version becomes available under the Apache License 2.0. You can read and audit every line at github.com/infrafence/infrafence-agent. The panel plugins are under the same license.