Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix [Bug]: Runtime context message leaked into assistant public reply and persisted to downstream LCM [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#73906•Fetched 2026-04-29 06:13:26
View on GitHub
Comments
1
Participants
2
Timeline
5
Reactions
0
Author
RayySummersRayySummers
Participants
clawsweeper[bot]clawsweeper[bot]RayySummersRayySummers
Timeline (top)
labeled ×2closed ×1commented ×1cross-referenced ×1

Runtime context text appeared in assistant's public reply instead of remaining internal, and downstream LCM persisted it as a normal conversation message.

Root Cause

This bug affects the downstream LCM plugin: Lossless-Claw has already filed a parallel report and added a defensive filter to reject messages starting with "OpenClaw runtime context". The root cause fix must come from OpenClaw core to prevent the runtime context from entering the assistant response stream in the first place.

Code Example

LCM assemble-debug log (conversation=85, session=5fd706df-e1ec-4266-ad9f-e4c9387ea2b2):

previousDivergenceMessage=assistant|blocks=1:text|text="OpenClaw runtime context for the immediately preceding user message. This context is runtime-generated, not user-author. Keep internal details private."

Full log entry timestamp: 2026-04-29T08:24:59.863+08:00
RAW_BUFFERClick to expand / collapse

Bug type

Behavior bug (incorrect output/state without crash)

Beta release blocker

No

Summary

Runtime context text appeared in assistant's public reply instead of remaining internal, and downstream LCM persisted it as a normal conversation message.

Steps to reproduce

NOT_ENOUGH_INFO

Expected behavior

Runtime context messages must remain internal and never appear in assistant's public response output, nor be persisted to conversation history.

Actual behavior

The internal runtime context text appeared in the assistant's public reply. LCM logs confirm it was stored as a normal assistant message (previousDivergenceMessage=assistant|blocks=1:text|text="OpenClaw runtime context for the immediately ..."), causing divergence on subsequent turns.

OpenClaw version

2026.4.26

Operating system

Windows 11 with WSL2 (Ubuntu 22.04)

Install method

No response

Model

Primary: MiniMax-M2.7

Provider / routing chain

openclaw -> minimax

Additional provider/model setup details

No response

Logs, screenshots, and evidence

LCM assemble-debug log (conversation=85, session=5fd706df-e1ec-4266-ad9f-e4c9387ea2b2):

previousDivergenceMessage=assistant|blocks=1:text|text="OpenClaw runtime context for the immediately preceding user message. This context is runtime-generated, not user-author. Keep internal details private."

Full log entry timestamp: 2026-04-29T08:24:59.863+08:00

Impact and severity

Affected users: End users of OpenClaw-powered bots Severity: Annoying (internal text leaks to user) + blocks workflow (LCM history polluted) Frequency: Observed once; trigger conditions unknown Consequence: Users see internal system text; conversation session requires manual reset to recover normal behavior.

Additional information

This bug affects the downstream LCM plugin: Lossless-Claw has already filed a parallel report and added a defensive filter to reject messages starting with "OpenClaw runtime context". The root cause fix must come from OpenClaw core to prevent the runtime context from entering the assistant response stream in the first place.

extent analysis

TL;DR

The most likely fix involves modifying the OpenClaw core to prevent runtime context text from being included in the assistant's public response.

Guidance

  • Review the OpenClaw core code to identify where the runtime context text is being added to the assistant's response stream.
  • Verify that the previousDivergenceMessage log entry is correctly formatted and that the text field does not contain internal runtime context.
  • Investigate the openclaw -> minimax provider/model setup to ensure that it is not contributing to the issue.
  • Consider adding a filter or check in the OpenClaw core to prevent internal runtime context text from being sent to the LCM plugin.

Example

No code snippet is provided as the issue does not contain enough information about the specific code implementation.

Notes

The root cause of the issue is unclear, but it appears to be related to the OpenClaw core. The Lossless-Claw defensive filter may help mitigate the issue, but a permanent fix is needed to prevent the runtime context text from entering the assistant response stream.

Recommendation

Apply a workaround by adding a filter to the OpenClaw core to prevent internal runtime context text from being sent to the LCM plugin, until a permanent fix can be implemented.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

FAQ

Expected behavior

Runtime context messages must remain internal and never appear in assistant's public response output, nor be persisted to conversation history.

#device allocation#model download#tokenizer error#prompt formatting#conversation history

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix [Bug]: Runtime context message leaked into assistant public reply and persisted to downstream LCM [1 comments, 2 participants]