fix(dashboard): mode toggle survives the auto-refresh; installed leve…

[contactacefleetdev]

…rs are visible

Two bugs, both reported from actually using the page.

THE TOGGLE SNAPPED BACK TO OFF. Not the handler -- a programmatic click set shadow and held. The page carries , so it reloads itself every 20 seconds and the mode, being client-side state with nowhere to live, reset. It reads as a broken control rather than as the page refreshing, which is exactly how it was reported. The selection now persists in localStorage and is restored on load: a per-viewer display preference, nothing another viewer needs, and it survives the reload that caused the problem. Reads and writes are wrapped -- a private window or blocked site data returns 'off' rather than throwing.

AN INSTALLED LEVER WAS INVISIBLE. Discovery found all three, the payload carried them, every mode resolved them -- and the rail rendered only the simulated headroom rows and the measured ones, so installed was in the payload and read by nothing. A developer who installed a lever package saw no evidence anywhere on the page that it had worked.

They now render as their own group, carrying no figure. A registered lever that is off has produced neither a measurement nor a simulation, and inventing a number would put a third claim on a rail that already has to keep two apart. What it shows is what is actually actionable: the providing package, the declared risk, whether the lever needs the proxy path, and the one word standing between it and doing something.

Also fixed on the empty-scope path, which returned early and skipped both the measured and installed rows -- whether a package registered a lever has nothing to do with whether there are sessions to score it against, so an empty range was hiding the only evidence the package was present.

Description

Summarize the changes made in this Pull Request and the problem being solved.

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update

Verification

Describe the tests and checks performed to verify changes:

• pytest passes cleanly.
• python -m build and twine check dist/* pass.
• E2E test script scripts/e2e_test.py passes.

Checklist

• My code follows the code style of this project.
• I have updated documentation where necessary.
• I have updated CHANGELOG.md with my changes.