dsh-rice-patrol

Rice Patrol 🍚🐋

简体中文

Rice Patrol is a DeepSeek Harness (DSH) plugin for repetitive reasoning, which we call “singing.” It reads visible reasoning text through DSH's general stream interface, without a provider-specific transport.

How detection works

The plugin splits reasoning text into non-empty, newline-terminated lines, trims whitespace at each end, and watches the latest 50 lines. It raises a warning when their median length is at most 20 characters and at least 70% of the lines have an exact duplicate in that window. If another qualifying window appears after at least 50 more lines, it confirms repetition.

It does not look for phrases such as Let me, require lines to get progressively shorter, or stop at the first warning. Each model request is checked separately. Repetition in ordinary answer text, and reasoning DSH cannot see, are outside its scope.

What happens after confirmation

Choose a mode in Plugins → rice-patrol:

Mode	Behavior
Stop (default)	Stop the current task and wait for you.
Observe	Record the signal without stopping the task.
Recover	Stop the task, compact verified work records, and let one fresh Agent try to continue.

Recover makes at most one automatic attempt. If it cannot verify a background job or tool result, it pauses for you instead of guessing. Finished or blocked reminders have a dismiss button; their records remain. Buttons and explanations follow the DSH interface language (Chinese or English). When no language is explicitly selected in DSH, it uses the browser's system language preference. Changing mode requires a DSH Web restart and does not change an in-flight task.

The handoff checks tools and background jobs first, extracts a clean record without the degenerate reasoning tail, compacts that clean record once, then starts a fresh Agent. If that one compaction reaches its output token limit, Rice Patrol discards the partial summary and continues from verified user and operation facts only. Timeouts, provider errors and unfinished streams still pause recovery. Uploaded user images are checked and passed to the new Agent through DSH's saved image references; the compaction call receives no image bytes. It does not compact the full old session or delete completed file edits.

Install

Tested with DSH 0.1.6-alpha.2:

dsh plugin --profile web add github:LeeLizuoLiu/rice-patrol#v0.2.5

Restart DSH Web, enable both rice-patrol and its component on the Plugins page, then choose a mode in its settings. Installing the package alone does not activate the guard.

Limits and possible next steps

• The 50-line window and two qualifying windows are fixed today. A future update may let users choose the lines per window and the number of windows needed for confirmation.
• Legitimate tool handoffs have appeared about two seconds after confirmation. Stop and Recover try to cancel immediately and may interrupt such work. Use Observe to assess your workflow first.
• Recover makes extra model calls. Its limits are 45 seconds for compaction and, for the new Agent, 5 minutes, 16 model requests, and 128 tool calls. Local cancellation does not prove provider billing has stopped.
• Behavior still needs validation across models and providers. A plain Output token limit reached message is not a detection signal.
• Recover still pauses if a user attachment is missing or uses an unsupported format, including file attachments. A previous paused recovery is not retried automatically.

v0.2.5 adds a bounded fallback for token-capped compaction and clearer pause messages. The WeakDALearning incident was inspected read-only; its exact compaction finish reason was not retained, so this change addresses a plausible cause rather than claiming to prove it. 60 local tests passed, including a synthetic AgentLoop fallback. No real model call was made for this fix. A previously paused recovery will not restart automatically. See the recovery validation and earlier Web validation for test scope and technical details. MIT licensed.