Tel-Agent answers phone calls, holds provider API keys, and can place outbound calls that cost real money. Security reports are taken seriously.

Do not open a public issue, pull request, or discussion for a security problem.

Report it privately, either way:

1. GitHub private vulnerability reporting — the Security tab of this repository, "Report a vulnerability". Preferred: it keeps the report attached to the code.
2. Email — info@dpro.at with SECURITY in the subject.

Please include what you can: what the issue is, how to reproduce it, what an attacker gains, and the version or commit you tested.

What to expect. This is a small team on a young project, so no formal response time is promised. We aim to acknowledge a report within five working days and to keep you informed while it is being looked at. If you have heard nothing after that, send a reminder — it means the message was missed, not ignored.

You will be credited in the fix unless you prefer otherwise.

There is no released version yet. Tel-Agent is at Milestone 0 and is not ready for production use. Do not run it on a business phone line.

• Anything that lets an unauthorised party place an outbound call. There are three such paths — /api/calls/outbound, /hooks/call and the MCP endpoint — and toll fraud on any of them costs the operator money directly. See §B9.1 of the specification.
• Exposure of API keys or SIP credentials, in logs, API responses, error messages, or the interface.
• Unauthorised access to call recordings or transcripts. These are personal data under GDPR.
• Authentication and authorisation flaws, including anything that lets one tenant read another's data.
• Bypassing the recording announcement. In Austria both parties must be aware a call is recorded, so suppressing it is a legal problem, not only a technical one.
• Injection through call content — a caller who can make the agent perform an action it should not, for example invoking a tool through what they say.

• Deployment choices the operator makes. Exposing the port to the internet, reusing a weak SIP password, or skipping HTTPS are documented risks, not vulnerabilities in Tel-Agent. The documentation states the supported paths: VPN, Tailscale, or a reverse proxy with HTTPS.
• Vulnerabilities in third-party providers. Report those to the provider. Tell us as well if Tel-Agent's use of them makes the impact worse.
• Missing hardening headers or similar findings with no demonstrated impact.
• Anything requiring an attacker to already have administrative access.

Until there is a release, these matter more than any code fix:

• Do not expose the service to the internet. Use a VPN, Tailscale, or a reverse proxy with HTTPS.
• Configure geographic dialling permissions at your telephony provider and disable every country you do not need to call. This is the single most effective protection against toll fraud, it is free, and it works even if Tel-Agent itself has a bug. It is not enabled by default on a new provider account.
• Use a strong password on the SIP extension. Brute-forcing weak extension credentials is a common way in.
• Disable outbound calling on the agent's extension while the agent only answers calls.
• Keep .env out of version control. It is gitignored; keep it that way.