Security
How to report a security problem in RDMBench, what to send, and what happens after you send it.
RDMBench runs on the bench computer, talks to fixtures on a DMX line or an Art-Net network, and has a built-in MCP server whose tools can drive a fixture. If you have found a way to make any of that do something it should not, please say so privately first.
Where to send it
Email darin@rdmbench.com. Please don’t post it in public or open an issue about it until a fixed version is out.
The same address takes ordinary problem reports: something RDMBench itself got wrong, rather than the fixture.
What to include
- The RDMBench version and the computer’s system. About RDMBench shows both.
- What you did, what happened, and what should have happened instead.
- Whether it needs anything on the network, a fixture on the line, or an assistant connected.
What counts
Anything that lets RDMBench be made to:
- send RDM or DMX to a fixture that nobody asked it to, or keep output up after it says it has ended;
- let another computer reach the MCP server without its password, or let an assistant do more than the server’s tools allow;
- read or write files outside its own folders and the ones you pick;
- install an update that RDMBench did not sign, or a fixture library that is not signed;
- leak the GDTF Share password, or anything else it keeps in the system’s secret store.
A fixture that breaks the RDM standard is not a security problem in RDMBench. The maker should hear about it. RDMBench’s behavior check gives you the packets to send them.
What happens next
A person reads every report and replies to it. A fix ships as a new version on every platform at once, and its release notes say a security problem was fixed, with your name if you want it there. There is no bug bounty: RDMBench is free, and so is the work of reporting.
For agents, the same contact is in
/.well-known/security.txt, in the
standard format (RFC 9116).