User Reports Claude Code Feature Sending Erroneous Messages Across Sessions

A user says a newly added Claude Code feature frequently sends messages to other active sessions, often incorrectly and with little practical value. The feature is reportedly enabled by default, but its disable setting remains unclear.

Published 9/10/2026, 4:27:23 AMLast updated 9/10/2026, 4:27:23 AMEditorial publication

A user has reported that the latest Claude Code release introduced a feature capable of sending messages to other running sessions. According to the post, the feature often makes incorrect decisions about when to send a message, creating distraction and consuming additional tokens.

The claim is currently supported mainly by the original social-media post and a reposted account. No clear version number, official changelog entry, or reproducible test has been provided. A follow-up post appears to point toward a possible setting, but the available material does not establish the exact control or whether it fully disables the behavior.

Other material included with the report discusses adaptive thinking, default effort levels, usage analysis, and a separate reasoning-history bug. Those reports do not directly verify the cross-session messaging claim, so this should be treated as an unconfirmed user complaint rather than an established product defect.

Source evidence 6

Source review does not guarantee permanent accuracy. Consider the original source, observation time, and later corrections.

An update on recent Claude Code quality reportsanthropic.com · supporting

The implementation had a bug. Instead of clearing thinking history once, it cleared it on every turn for the rest of the session. After a session crossed the idle threshold once, each request for the rest of that process told the API to keep only the most recent block of reasoning and discard everything before it. This compounded: if you sent a follow-up message while Claude was in the middle of a tool use, that started a new turn under the broken flag, so even the reasoning from the current turn was dropped. Claude would continue executing, but increasingly without memory of why it had chosen to do what it was doing. This surfaced as the forgetfulness, repetition, and odd tool choices people reported. [...] As part of the investigation, we back-tested Code Review against the offending pull requests using Opus 4.7. When provided the code repositories necessary to gather complete context, Opus 4.7 found the bug, while Opus 4.6 didn't. To prevent this from happening again, we are now lan