Synapse is an infrastructure component for organizations of AI agents. It handles credentials, messages and task data — security is treated as a first-class concern.

Please do not open a public issue for security vulnerabilities. Report them privately via GitHub's Security Advisories:

1. Go to the repository's Security tab;
2. Click Report a vulnerability;
3. Describe the vulnerability, the impact, and a minimal reproduction.

You can also email the maintainers directly (see the repository profile for the contact address).

We aim to acknowledge reports within 48 hours and to provide a first assessment within 7 days. We keep reporters informed until a fix is released. If the vulnerability is accepted, a fix is prepared privately and released with an advisory; reporters are credited unless they prefer to stay anonymous.

• Credentials are never stored in plain sight: initial passwords are kept in root-only files, secrets are passed via stdin, never via command-line arguments or environment variables.
• The server listens on a Unix socket with restrictive permissions and on the loopback interface only — no public network exposure by design.
• All writes go through SQLite transactions; backups are encrypted and verifiable (restore proof).
• Every command is authenticated and authorized individually; a standard account cannot escalate privileges, and organizations are permanent (no deletion path).
• Systemd units enforce memory limits, filesystem restrictions and explicit directory permissions.

If you find a way around any of these, we want to hear about it — privately.