v1.5.5

Debug visibility. Events tab and a new Conditions tab carry the
"why is this resource sad" story for every workload kind — no more
jumping back to kubectl describe. Events aggregation now walks the
full controller chain (CronJob → Jobs → Pods, Deployment → current
ReplicaSet → Pods, StatefulSet/DaemonSet/Job → Pods). Panel 2 menu
gains Enter (drill) and Esc (back) entries for discoverability. Logs
auto-follow indicator switched from a ▼ marker to coloring the active
[Logs] label green.

Added

• Conditions tab. New detail-panel tab showing the resource's
  .status.conditions as a TYPE / STATUS / REASON / MESSAGE / AGE
  table — same content as the Conditions section of kubectl describe.
  Status False rows highlighted in red. Tab appears for kinds that
  populate conditions: Pod / Node / PVC / Deployment / StatefulSet /
  DaemonSet / Job / HorizontalPodAutoscaler / Ingress. Hidden for
  kinds without conditions (ConfigMap, Secret, Service, ServiceAccount,
  etc.). Why it matters: events expire after the cluster's TTL
  (default 1h), but conditions reflect the resource's current state
  — PodScheduled: False, Insufficient cpu stays visible until the
  Pod is actually scheduled.
• Aggregate child events for all workload kinds. Selecting a
  workload's row and switching to the Events tab now merges events
  from the workload itself AND its child Pods, sorted newest first.
  The Object column distinguishes source kind so you see the chain
  inline. Covers Deployment (via current ReplicaSet), StatefulSet,
  DaemonSet, Job, ReplicaSet, CronJob. Mirrors the existing aggregate-
  logs pattern; same PodsForWorkload helper drives both.
• CronJob 3-tier aggregate. CronJob's Events tab additionally
  pulls in events from every owned Job (so you see Job-level
  SuccessfulCreate / BackoffLimitExceeded / Completed alongside
  the CronJob's SuccessfulCreate / MissingJob and the Pods'
  Scheduled / Pulled / Started). Three layers in one view —
  the killer feature for "why did last night's cron fail" debug.
• Conditions tab Space hint. Same cheatsheet pattern as Logs /
  Events tabs — j/k/u/d/gg/G/y/z for scroll/copy/expand.
• Panel 2 menu Enter (drill) entry. When the selected kind
  supports drill-down, the per-row Space menu now ends with an
  Enter ↘ row whose hint names the child kind ("drill into pods" /
  "drill into containers" / etc.). Cursor + Enter on it triggers
  the same drill as pressing Enter on the row directly — visible
  in the menu for discoverability.
• Panel 2 menu Esc (back) entry. When the table is inside a drill
  chain (e.g. you pressed Enter on a Deployment and are now viewing
  its Pods), the menu appends an Esc ↖ back to parent list row.
  Cursor + Enter triggers exitDrillDown.
• Container menu Esc entry. Same back row appended to the Pod →
  containers context menu (Shell + Esc).
• Panel 3 bottom-left hint. On the Relatives tab at depth > 1,
  panel 3's border shows esc: back in the bottom-left, mirroring
  panel 2's .: toggle helm pattern. Surfaces the pop-one-level
  shortcut without requiring the help popup.

Changed

• Logs follow indicator. Active [Logs] tab label rendered in
  Status.Running green when auto-follow is engaged. Replaces the
  prior ▼ text marker. Same semantic ("alive stream") expressed
  visually instead of textually.
• Popup bottom hints trimmed. All popup border bottom legends
  (panel 2 menu, helm doc menu, breadcrumb, hint popup, app log,
  YAML popup, confirm, namespace picker, context picker) now show
  just Space: close / Space: cancel. The Esc / q / n / ! keys
  still work, just no longer advertised. Reflects the v1.5.x
  mental model: Space is the primary popup verb.
• CronJob added to demo fixtures. .local/demos/demo-app.yaml
  now includes a demo-cron CronJob firing every minute (busybox
  heartbeat) for verifying the 3-tier aggregate path locally.

Fixed

• Stale-events workaround framing. Empty Events tab on a healthy
  resource is no longer the only signal — the new Conditions tab
  fills the diagnostic gap when events have expired past TTL.