Audit Workflows and Schedule polish found during Task

Seen while screenshotting the Audit screens for Task 7132: 1. Audit › Workflows › Schedule at ≤480px: the skip-reason pill (`already_running`) is gated off on the phone, and nothing on the row replaces it. Hang the reason, or fold it into the mark's title, so it can still be read on a phone. 2. Audit › Workflows › Runs at desktop: the run-id column still ellipsises ids like `task-notify:44:system_f…` while the reading column has hundreds of spare pixels. Size the run id to its content, the way the Workflow column is now sized (`contentSlotWidth`). 3. Audit › Sessions › Efficiency at desktop: with analytics unavailable, two identical 7d/14d/30d pickers are stacked one above the other (the page's and Project memory's). Consider sharing one period, or labelling which one each picker drives. 4. `pages/cron/vocabulary.ts::detailLines` crashes the whole Schedule tab when a firing's `detail` is null. The type says string, but one bad row should not take down the page. Treat null as empty. Out of scope: API data shapes, except as item 4 needs. Done means: - Each item shipped with a test, or declined with a reason. - Visual check at desktop and ≤480px. - `./dc check full` green.

All 4 items shipped with tests: phone skip reason hung, run-id sized to content, one shared Efficiency period, null detail safe. PR #5846 merged (d77a7d3, full gate green); visual check done at 1440/400. Its deploy was superseded by a later push. Follow-up draft #7198.