InstanceState Reload can cause client to hang on subagent return

[RainbowLazer]

Description

Debugged the issue source live while running OpenChamber, using Claude code. As I understand it, if InstanceState refreshes during a subagent task, on return to the parent agent, the old state is retained in the cache and fails to match with the new state, causing a deadlock. In sessions with many subagents, or long-poorly-timed background tasks, this can cause OpenCode to become unresponsive to anything on subagent return until restarted.

Plugins

No response

OpenCode version

No response

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response

[github-actions]

This issue does not fully meet our contributing guidelines.

What needs to be fixed:

• Steps to reproduce are missing (_No response_). Please describe how to trigger the hang — e.g., the type of session, number/depth of subagents, and what actions lead to the unresponsive state.
• OpenCode version is missing. Please provide the version you are running (opencode --version).
• Operating System and Terminal are missing. These help narrow down platform-specific behavior.

Please edit this issue to address the above within 2 hours, or it will be automatically closed.

---

This issue might be a duplicate of or closely related to existing issues. Please check:

• Stale per-directory provider state after external auth.json change: ProviderModelNotFoundError with misleading "Did you mean" until POST /global/dispose #50647: Stale InstanceState per-directory provider cache causes failures after external auth changes — same InstanceState caching mechanism, different trigger and symptom.

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

This issue has been automatically closed because it was not updated to meet our contributing guidelines within the 2-hour window.

Feel free to open a new issue that follows our issue templates.