[BUG] Remote trigger routes second automation fire into still-running session instead of spawning a new one

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

Bug description
When a Monday.com automation fires the Claude Code remote trigger twice in quick succession (two separate remote_trigger API calls, each with a distinct payload), the second call was delivered as a new user message into the already-running first session rather than spawning an independent session.

Expected behaviour
Each remote trigger call → new isolated session, own branch, own PR. The two jobs are unrelated and should never share a context window or a git branch.

Actual behaviour
Both trigger payloads arrived as sequential user messages inside session fcbeabe3-22c1-5f7b-9495-cd4138a9c3f4. The session remained active after completing the first task, and the second trigger was routed into it. Because the harness assigns one branch per session (claude/serene-knuth-lzx1v2), both commits landed on the same branch and were included in the same PR — mixing two unrelated changes from two different requesters.

Reproduction
Configure a Monday.com automation to call the Claude Code remote trigger when a pulse status changes to a specific value.
Change two unrelated pulses to that status within a short window (both while the first session is still running).
Observe: second trigger payload arrives as a new message in the first session rather than a fresh session.

What Should Happen?

Expected behaviour
Each remote trigger call → new isolated session, own branch, own PR. The two jobs are unrelated and should never share a context window or a git branch.

Error Messages/Logs

Steps to Reproduce

Reproduction
Configure a Monday.com automation to call the Claude Code remote trigger when a pulse status changes to a specific value.
Change two unrelated pulses to that status within a short window (both while the first session is still running).
Observe: second trigger payload arrives as a new message in the first session rather than a fresh session.

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

it is in the web

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_