Continuous operation + live tracking on the central app

Max: 'running continuously at all times' + 'a central app for tracking them all, including those in progress.' Not a cron sweep — an always-on lane pool:

1. Triggers: new external release (source watchers), new certified Populace artifact, engine version bump — each re-scores affected comparisons; full-corpus re-score gates certification (epic #1 point 3).
2. Loop-until-dry: lanes drain the worklist; when dry, watchers idle until a trigger fires. Lane dispatch per Max's routing (sol grind behind deterministic comparison gates; fable adjudication on publish).
3. Live status feed: every lane writes status transitions (registered → cataloged → ingested → computed → diagnosing → published → regressed) to a feed the app renders — in-progress work visible, not just results. Options: repo-committed status JSON (simple, auditable) vs Supabase telemetry (populace pattern, live). Recommend starting repo-committed, graduating to Supabase when lane count makes commits noisy.
4. The app's home view = mission control: corpus coverage, running lanes, freshest divergences, held-out win/miss record.

No files or tests are named. Start by mapping the source watchers, certification and engine-version triggers, lane worklist, status feed, and central app described here; done means continuous draining, visible in-progress transitions, and mission-control views for coverage and outcomes.