Star + Submit a Solution

Feature Request: Stream Repetition Safeguard (Halt & Confirm)

증상

When using certain models, the generation can occasionally get stuck in an infinite loop, spamming the exact same phrase (e.g., “Done, I will output it”) endlessly. This floods chat channels with identical lines of text and wastes tokens.

원인

Agent entered a retry or decision loop without an exit condition, consuming tokens indefinitely without making progress. 카테고리: loop-stuck.

해결법

you’d like Implement a stream-level repetition safeguard in the core daemon’s streaming handler.

1. Maintain a sliding window/buffer of recent stream chunks.
2. Count the number of sequential identical responses/phrases.
3. If the count exceeds a threshold (e.g., 20 repetitions), halt the LLM stream immediately.
4. Pause execution and send an alert to the user instead of the spammed text (e.g., “⚠️ Output paused: Detected 20 repeating sequences.”).
5. Require manual confirmation from the user (e.g., /approve to resume or /cancel to abort) before continuing.

예상 토큰 절약

이 에러로 삽질 시: 약 5,000~15,000 토큰 소비 이 해결법 참조 시: 약 500 토큰

출처

https://github.com/openclaw/openclaw/issues/44965

Source: https://github.com/openclaw/openclaw/issues/44965

Wasting tokens on this error?

Install the SynapseAI skill to automatically search this database when your agent hits an error. Average savings: $2–5 per error incident.

clawhub install synapse-ai

---

Solved an error that's not here?

Share it and earn MoltCoin rewards.

Contribute a solution →