RDC failed in ChatGPT — what should I check first?
A blocked tool call, a connection problem and a local command error can look similar. Start with the safe troubleshooting guide; it helps you identify which kind of failure you are seeing.
Short answer
Use the guide below to check whether the issue is an OpenAI safety-check rejection, a connector or session problem, or a command/application error. A failure message alone does not establish the cause. The diagnostic steps do not bypass ChatGPT safety controls.
What the available test shows
One documented sample session recorded 20 successful results from 20 checks: 14 dedicated RDC operations and six exact file-backed Python/ROSE calls, with no explicit OpenAI safety-check rejection in that sample. This was one session, not a controlled trial. It does not show permanent reliability or prove intermittent safety-check rejections are fixed.
John has also reported approximately 15–30 deep investigations into ChatGPT / Remote Desktop Commander reliability. That is a personal report, not an independently audited count.
Which “RDC” is this about?
This troubleshooting page concerns the existing Remote Desktop Commander connector in ChatGPT. RDC by Rose is a separate Petal that remains in development; this page does not activate it or promise availability.