Summary
Hello everyone. After updating my AGENTS.md and subagents to have an orchestrator flow, I decided to use 5.6 Luna as my orchestrator, which I instructed on my AGENTS.md to use the selected model as the main one. I changed to 5.6 Luna, set effort to Medium, sent a message, and immediatly received: "⚠ Error: 400 [OpenAI] Invalid 'input[314].name': string does not match pattern. Expected a string that matches the pattern '^[a-zA-Z0-9_-]+$'.
Type "continue" to try again. If the issue persists, contact support: https://commandcode.ai/discord
Trace ID: 9079642d20d7bc5dab038c37ab160a41"
This is all I did, just change to the model and sent a message.
Update: Actually, don't know what it is, Luna actually works on a new session, but fails in the session that I had, no matter the effort set.
As you can see in the image (I cut out things from my code to not share), on this session, I mainly used DeepSeek V4 Flash without issues, once I swapped to MiniMax M3 to help me with a merge conflict, and then back to DS Flash to finish the implementation. I first tried Luna for the merge conflict, where I got the error, and as you can see here, change again to Luna and the error always happens, but again, only in this session, in a new one works fine.
One thing that I did, but I don't know if it's relevant, I started this session when I was in one branch, then mid session I swapped branches, don't know if this could have caused issues, probably not, but just to note.
Update 2: Actually, I don't think it's just a 5.6 Luna issue, when I was trying to export the session since something on it is causing the issue, I changed to ds flash and asked it to use the export command and then remove everything that might be sensitive data, and I got the exact same error again. Ending that instance, starting a new one and entering the same session again allowed me to continue, don't know really what the issue is.
Expected Behavior
GPT 5.6 Luna works without issues.
Actual Behavior
GPT 5.6 Luna immediately fails only in this session with error "Error: 400 [OpenAI] Invalid 'input[314].name': string does not match pattern. Expected a string that matches the pattern '^[a-zA-Z0-9_-]+$'."
Steps to reproduce the issue
I don't know, it only happens in the session I was in, but I didn't do anything out of the ordinary.
Command Code Version
1.28.1
Operating System
Windows
Terminal/IDE
Unknown
Shell
cmd.exe
Session file (optional)
No response
Fix prompt (optional)
No response
Additional context
No response
Summary
Hello everyone. After updating my AGENTS.md and subagents to have an orchestrator flow, I decided to use 5.6 Luna as my orchestrator, which I instructed on my AGENTS.md to use the selected model as the main one. I changed to 5.6 Luna, set effort to Medium, sent a message, and immediatly received: "⚠ Error: 400 [OpenAI] Invalid 'input[314].name': string does not match pattern. Expected a string that matches the pattern '^[a-zA-Z0-9_-]+$'.
Type "continue" to try again. If the issue persists, contact support: https://commandcode.ai/discord
Trace ID: 9079642d20d7bc5dab038c37ab160a41"
This is all I did, just change to the model and sent a message.
Update: Actually, don't know what it is, Luna actually works on a new session, but fails in the session that I had, no matter the effort set.
As you can see in the image (I cut out things from my code to not share), on this session, I mainly used DeepSeek V4 Flash without issues, once I swapped to MiniMax M3 to help me with a merge conflict, and then back to DS Flash to finish the implementation. I first tried Luna for the merge conflict, where I got the error, and as you can see here, change again to Luna and the error always happens, but again, only in this session, in a new one works fine.
One thing that I did, but I don't know if it's relevant, I started this session when I was in one branch, then mid session I swapped branches, don't know if this could have caused issues, probably not, but just to note.
Update 2: Actually, I don't think it's just a 5.6 Luna issue, when I was trying to export the session since something on it is causing the issue, I changed to ds flash and asked it to use the export command and then remove everything that might be sensitive data, and I got the exact same error again. Ending that instance, starting a new one and entering the same session again allowed me to continue, don't know really what the issue is.
Expected Behavior
GPT 5.6 Luna works without issues.
Actual Behavior
GPT 5.6 Luna immediately fails only in this session with error "Error: 400 [OpenAI] Invalid 'input[314].name': string does not match pattern. Expected a string that matches the pattern '^[a-zA-Z0-9_-]+$'."
Steps to reproduce the issue
I don't know, it only happens in the session I was in, but I didn't do anything out of the ordinary.
Command Code Version
1.28.1
Operating System
Windows
Terminal/IDE
Unknown
Shell
cmd.exe
Session file (optional)
No response
Fix prompt (optional)
No response
Additional context
No response