FE-1724: Remove the model-facing Brunch question marker

[kostandinang]

🌟 What is the purpose of this PR?

Remove the model-facing question replay marker so direct questions no longer require a marker tool call and extra model completion.

🔗 Related links

• FE-1724 (internal)
• Stack parent #9704

🚫 Blocked by

• #9704

🔍 What does this change?

• Stops mounting, prompting, and cataloguing both question-marker tool names.
• Preserves legacy marker parsing and hiding for existing conversation history.
• Verifies new questions complete in one provider call without marker data while full-response playback remains available.

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

• does not modify any publishable blocks or libraries, or modifications do not need publishing

📜 Does this require a change to the docs?

The changes in this PR:

• require changes to docs which are made as part of this PR

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

• do not affect the execution graph

⚠️ Known issues

New responses intentionally do not offer “Repeat question”; historical marked questions remain compatible.

🐾 Next steps

None.

🛡 What tests cover this?

• Core and transport unit tests for absent mounting and legacy projection.
• Website Voice tests for full-response speech without a repeatable question segment.
• Brunch integration tests for a marker-free, single-completion question turn.

❓ How to test this?

1. Run the affected unit and integration suites.
2. Ask a direct question through Brunch.
3. Confirm the response completes normally without a marker tool call and remains available to “Read full response”.

📹 Demo

Not applicable; this removes internal tool orchestration without changing the rendered response.

Made with Cursor

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
hash	Ready	Preview	Sep 15, 2026 3:10pm UTC
petrinaut	Ready	Preview	Sep 15, 2026 3:10pm UTC
petrinaut-docs	Ready	Preview	Sep 15, 2026 3:10pm UTC

1 Skipped Deployment

Project	Deployment	Actions	Updated
hashdotdesign-tokens	Ignored	Preview	Sep 15, 2026 3:10pm UTC

[cursor]

Stale Bugbot comment from a previous run.

[kostandinang]

GitHub locks base changes for PRs in a native stack. Closing this copy so FE-1724 can be recreated directly against main without disturbing the queued Voice stack.

[kostandinang]

Superseded by #9742, recreated directly against main. The original review fix is included in the rebased branch.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 8173c53. Configure here.}

[cursor]

Sweep keeps host-only success requirement

Medium Severity

requireSuccessfulRuns is fixed to true for the whole sweep session when a host request creates the experiment. After the assistant captures its result, requestActive is cleared and users are allowed to explore, but any later batch with a failed run still errors the experiment. User-created sweeps do not behave this way, so inspecting an AI optimization after completion can lock the experiment in an error state.

Additional Locations (1)

• libs/@hashintel/petrinaut/src/react/experiments/sweep-session.ts#L569-L578

 

^{Reviewed by Cursor Bugbot for commit 8173c53. Configure here.}