[Duang777]

This proposal is absorbed into #7209 at 35e99bf, retaining its separate signed commit. Closing this duplicate proposal lets one branch qualify both current-main CI failures.

Summary

The research-history smoke assumes one Escape closes analysis, but an open context-help popover correctly consumes Escape first. Close analysis explicitly and wait for the reader to disappear before testing an unanalysed scope.

- Escape, then select another scope
+ Close analysis, wait for it to disappear, then select another scope

Evidence

• Before: the complete research-history browser scenario fails with the analysis reader still open. A read-only DOM observation confirms that Escape closed the visible analysis-scope help popover and preserved the parent reader, as intended.
• After: the complete research-history scenario passes in both development and packaged builds. The later zero-analysis-write assertion is unchanged; existing help-dismissal and keyboard-close checks remain.

Merge Danger

Door: two-way

Blast Radius: tests

Only two smoke lines change. The application's Escape priority and analysis behavior are unchanged.

Goal And Delivered Outcome

• Outcome basis / optional anchor: repeated dashboard-browser (2) failure while following fix(bootstrap): validate goal ids before path construction #7192, fix(leases): canonicalize equivalent write scopes #7204 and fix(ci): restore digest ownership and browser smoke preconditions #7209.
• Goal/source and gap: a source-switching assertion starts before the test has closed its analysis reader.
• Observable before -> after: the same complete browser scenario fails before this change and passes after it, retaining assertions that scope selection neither opens analysis nor invokes it.
• Issue/task and intended base: related to fix(bootstrap): validate goal ids before path construction #7192, fix(leases): canonicalize equivalent write scopes #7204 and fix(ci): restore digest ownership and browser smoke preconditions #7209. Base: main at 659c1fd.

Author Declaration

• Written by: Codex, OpenAI, with user-authorized follow-up and submission.

Implemented against

• Specification and revision: existing research-history.mjs no-analysis-write and help-dismissal assertions at 659c1fd.

Criterion	Disposition	Symbol / path	Test or command
Explicitly close the analysis reader before switching factual scopes	implemented	research-history.mjs	Full research-history browser scenario
Selecting an unanalysed scope does not open or execute analysis	implemented	Existing unchanged assertions	Full research-history browser scenario
Dismissing help preserves the parent reader	implemented	Existing unchanged keyboard checks	Full research-history browser scenario

• Self-check before submission: reproduced the failure, observed the actual open popover and unchanged analysis route, inspected the Escape handlers, then ran the same complete scenario in both build modes.

Scope And Continuation

• Completed scope and remaining work: complete within the smoke's source-switching setup; remote CI remains pending.
• Slice boundary / successor: N/A. The separate packaged goal-hierarchy timeout was not reproduced in either build mode and its old failed build was queued for recheck.

Validation

• Tested revision: a85c4e72cbb79a25d8b0f2620754e8995b3753a2.
• Run state: finished.
• Input classes: synthetic.

Check kind	Result	Public-safe evidence / limitation
regression_parity	passed	Same complete research-history scenario fails before and passes after, without changing expected values or dropping the zero-analysis-write assertion.
integration	passed	Research-history scenario in development and packaged builds on Node 24.15.0 with an isolated Chromium context and synthetic fixtures. No live model or user data is used.
static	passed	Dashboard build, standard diff canary, bilingual diff check and selected public-boundary scan.
integration	not_run	Remote required CI has not yet run on this branch. Other browser scenarios were not rerun except goal-hierarchy, which passed in both modes.

• Coverage and gaps: this proves the smoke's reader-close precondition and unchanged assertions through a real browser, not backend analysis dispatch, model utility or production adoption.

Frontend / Visual Evidence

• UI impact: none.
• Before: N/A.
• After: N/A.
• First viewport and preview: N/A. Application rendering is unchanged; failure/pass and DOM observation evidence explain this test-only change.
• States and viewports shown: N/A.
• Source data: synthetic.
• Attention review: N/A.
• Craft review: N/A.

Type of Change

• Bug fix
• Test update

LoopX Area

• Build, packaging, installer, or CI

Technical Direction

• Direction / acceptance reference, when applicable: reliable browser qualification.

Shared-authority RFC fixture impact

• N/A. No provider, fixture schema, compatibility projection or authority policy changes.

Boundary Checklist

• No private state, credentials, raw traces, internal links or local paths are published.
• No maintainer-owned benchmark work is duplicated.
• The change stays scoped to the reproduced smoke failure.
• No paired RFC changed; bilingual check passed.
• UI impact is none.
• The commit includes a DCO Signed-off-by trailer.