The first DeepSeek Flash response works, but the request after a tool call asks for the previous thinking content. Where should I look?
The same screenshot works in my provider’s web app. In DeepSeek Harness, the attachment is sometimes blocked immediately and sometimes rejected after I send it. How can I tell where it fails?
Short chats work with self-hosted vLLM, while long chats and compaction fail. Which budget should I check?
Some requests work and others fail. How do I separate network, provider, and session problems?
The browser can open the endpoint, but Harness fails with SELFSIGNEDCERTINCHAIN. What should I check?
The same task produces weaker results through Harness. How do I isolate the difference?
The provider shows an unexpected model list and sometimes returns MissingSessionID. Are these the same issue?
Harness shows a reasoning setting. How can I check whether the server actually uses it?
Ollama is running, but Harness cannot connect or find the model. What should I enter?
I configured an API key and want to return to DeepSeek account sign-in. Where is the switch?
