dsh-repetition-guard
DeepSeek Harness 插件：当模型在推理或可见文本中重复循环某个词（“wait wait wait” / “počkat počkat …”）时，提前中止模型流，使提供商停止生成，避免消耗不必要的令牌。
插件会安装到这里；不确定时保持 web。
npx -y @deepseek-ai/dsh plugin --profile web add dsh-repetition-guard@0.1.0Repetition Guard 以 dsh-repetition-guard 发布，当前版本为 0.1.0。Plugin Hub 会校验它的 manifest，并保存精确安装来源，便于复现安装结果。
English | Česky
Stops a model from "looping" during generation — when it starts repeating the
same word over and over (wait wait wait …, počkat počkat …), sometimes
hundreds or thousands of times, burning tokens for nothing.
The plugin hooks the llm/stream waterfall (DSH's official extension point
around every streaming model call) and watches both reasoning-delta chunks
(the thinking text) and text-delta chunks (the visible answer). As soon as
the tail of either stream contains 12 identical words in a row (threshold
configurable), it stops the stream.
Stopping the stream makes the adapter abort its fetch → the provider stops
generating further tokens. The rest behaves like a normal end (the assembler
treats finish = stop), so the turn ends cleanly with whatever was generated
before the loop.
tailChars characters (160), so it doesn't slow
the stream or grow memory.dsh plugin --profile web add dsh-repetition-guard
Restart the DSH session; it then guards every model call (including subagents).
config.threshold (default 12)config.tailChars (default 160)repetition-guard: stopped a repeated-word loop…MIT