dsh-fetch-timeouts
5 ★
DeepSeek Harness plugin: raise Node's HTTP timeouts process-wide so slow local models (Ollama, LM Studio) are not cut off at 5 minutes
Tools
dsh-llm-gate
2 ★
Per-provider concurrency gate for DeepSeek Harness model requests
Auth
dsh-model-pin
1 ★
No description
Auth
dsh-turn-doctor
1 ★
No description
Tools
dsh-context-budget
0 ★
DeepSeek Harness plugin: keep a local model's context at a size your GPU handles well (measured prefill speed, hard ceiling, early compaction)
Memory
dsh-logbook
0 ★
No description
Tools