[teamleaderleo]

#33389 repairs the Bring your history stack on #33215. First-run presentation now opens as a top page, independent of Home pane replacement, and records the once-only flag only after presentation succeeds. The failed launch had logged a page creation against an unknown pane; its log did not identify which page. The new route avoids that race.

The repair also disambiguates the Chief CLI JSON parser, fixes the two unknown-value lint expressions, restores the missing gallery id and four grid-expanded fixtures, and removes three pointing-hand cursors. The sticky footer uses the opaque surface token. Existing base test repairs are carried forward for hidden shimmer text, Session Info expectations, preview controls, font/scroll trace setup, and the six-way visual-audit split; coverage is preserved.

DEBUG verification uses an explicitly supplied regular ZIP resolved beneath HOME, after the normal native gesture check. Native Chromium input reaches its actual page window through AppKit; DOM waits have bounded observer cleanup and matching transport deadlines.

User request: “Find why the fresh-home launch at your repair head showed Home without the Bring your history page (routing bug in #33215, or harness isolation), fix it on #33389 if ours, also fix #33215's two lint errors, the missing-gallery failure and the four grid-expanded browser cases, then on a fleet DEV build verify: page shown once on fresh home, ChatGPT export zip imported through the page, Chief memory readback, Done state captured. Reply with head, results, captures; stop.” Do not merge.

Validation at 53dcae0:

• Fleet DEV build c6f2bd1e7f4cf3ce2132f88b passed. Artifact sha256:37d70153698841c9ae82adefd4481077a6ce11cce0ac40b0cfdaddcd4011f2ab.
• Fresh empty home/app support/defaults: actual Chromium history page painted, app up 30 seconds, no crashes. The smoke startup prerequisite accepts the underlying Home tab while the top history page has selection; the journey independently checks the real painted visible page.
• Actual page Choose export → Import imported four memories from a synthetic two-conversation ChatGPT ZIP. Bundled Chief memory search read back “My favourite editor is Helix.” and “My verification project is the Lavender Lighthouse.”
• Real Done state captured. Done closed the page. Same app/HOME/defaults restart had no history page, stayed alive, and quit cleanly. Both app exits were zero; no crash reports or uncaught exceptions.
• Fleet web bundles, formatting, lint (zero warnings/errors), types and protocol check passed. Run 41644704226 passed all 30 affected test files with zero failures, timeouts or retries. Gallery audit: 369/369 states; grid: 204 cases, including all four grid-expanded cases with zero measured deltas. The checked snapshot has the same tree as this head (767fc1deac4324dd2f4d409cc9c716fa67727674).
• Fleet native step c6a264b60416098d0806a631 passed all 14 tests across ImportHistoryPageProviderTests, PageWebViewDebugPressTests and HistoryFirstRunRoutingTests, with no test-runner crash.
• Independent source review: SHIP at 53dcae0, no findings; reviewer opened the Done capture.

Prior art carried for test fixtures: #32953, #32965, #32946, #33124, #32955, #32939 (section-order assertion only), #32940, #33135, #33185. These are repairs already on base, not new product behavior.

No merge requested or performed.

🤖 Generated with Codex

Agent session: agent session

[blacksmith-sh]

Found 10 test failures on Blacksmith runners:

Failures

Test	View Logs
test/accent-glass-invariants-4.test.ts/accent-glass-invariants-4.test	View Logs
test/accent-glass-invariants-4.test.ts/accent-glass-invariants-4.test	View Logs
test/diff-virtualization.test.ts/diff-virtualization.test	View Logs
test/diff-virtualization.test.ts/diff-virtualization.test	View Logs
test/diff-virtualization.test.ts/diff-virtualization.test	View Logs
test/diff-virtualization.test.ts/diff-virtualization.test	View Logs
test/message-preview-browser.test.ts/message-preview-browser.test	View Logs
test/message-preview-browser.test.ts/message-preview-browser.test	View Logs
test/session-info-width-stages.test.ts/session-info-width-stages.test	View Logs
test/session-info-width-stages.test.ts/session-info-width-stages.test	View Logs

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[cmux-bot]

cmux status _{(one comment, edited in place; each line is updated by the job that owns it)}

Passes: CI passes on 53dcae0ba4.

CI passes on 53dcae0ba4 (run 41644571364 attempt 1).

Written by scripts/ci/classify_failures.py (ci-failure-attribution.yml); signatures are its SIGNATURES table. A machine verdict is the runner's fault, not this PR's; yours means the failing file is one this PR changes, also red on main that main's latest full suite fails the same way, seen on other PRs that it failed on another pull request's run lately.

CI fast guards passes on 53dcae0ba4 (run). This covers only the fast guards, not CI: CI's result is ci-status.

[teamleaderleo]

Review (subagent claude-opus-5-5 medium): SHIP at 53dcae0

First-run routing shows the page once (first workspace, no saved windows; upgrades record the flag without opening it; HistoryFirstRunRoutingTests cover all three); no launch scan of Downloads/Desktop (the test-zip hook is DEBUG-only and HOME-bound with a symlink-escape test); Skip/Done and the Home memory offer unchanged; Swift thin. Fleet DEV build: fresh-home page shown, ChatGPT export imported through the page, Chief read back the imported facts, not shown again on restart.

Follow-ups: retry the offer if the first-run page fails to open.