Prompt processing suddenly stops (using LM Studio)

[jacobvalter]

Summary

I'm running 0.1.7 on middle-scale-size project using LM studio:

│ Provider LM Studio │
│ Model qwen3.5-35b-a3b │
│ Endpoint http://localhost:1234/v1 │

simple (short) prompts are working fine but when issued larger one, it suddenly stops in the middle of processing without any response.

context window has not been filled (using 256k on MBP M4 PRO with 48GB)

Steps to Reproduce

1. use LM studio / qwen3.5
2. use openclaude on large repo
3. issue prompt requiring heavy processing

Expected Behavior

prompt is processed with some output

Actual Behavior

response processing ends in the middle without any output

Environment

• OpenClaude version: 0.1.7
• OS: macOs 26.4
• Terminal: zsh
• Provider: LM Studio
• Model: qwen3.5-35b-a3b

Logs / Screenshots

...
I need to implement support for its reading. Beside this [repo name] repo there are other repos

• [repo name B] where usually these common/shared operations/code is placed. But specifically this
  DB schema implementation is made in [repo name C] repo. I believe moving it to [repo B]
  is the right thing.. but it uses EF core compared to existing implementation in [repo B].. can you
  judge and consult the right approach ?

⏺ I'll analyze your situation and provide guidance on the best approach for implementing SafetyNet
Curing ML database support in Tooling WebAPI.

Let me first explore the existing codebase to understand the current patterns and implementations.

Searched for 3 patterns (ctrl+o to expand)

⏺ Let me use Bash to explore the directory structure:

Listed 1 directory (ctrl+o to expand)

⏺ Good, I can see the directory structure. Let me explore the key repos:

Listed 3 directories (ctrl+o to expand)

✻ Brewed for 3m 21s

❯ and? what else? i haven't got any reponse

⏺ Let me explore the relevant repos to understand the current patterns and provide you with proper
guidance.

Listed 1 directory (ctrl+o to expand)

✻ Churned for 1m 14s

❯ hmm, churn more

Searched for 1 pattern, read 2 files (ctrl+o to expand)

✻ Cogitated for 2m 44s

❯ /exit
⎿ Catch you later!

[Vasanthdev2004]

Closing due to lack of detail (no version, LM Studio config, or model info). Prompt processing stops with local models can have many causes. If still happening, please open a new issue with reproduction details.