Skip to main content

KleinHub · Pulse

Mission control — operational awareness and system visibility

MONITOR

NewsStand recovery under observation

CASE-004 · Release case on NewsStand + University

Medium
Watching
Release
WatchingGuardian Waiting
Little Buddy — what this case meansplain language · explains only
Known

What happened: Prior warning state.

Known

Who it affects: NewsStand readers on NewsStand + University.

Known

Proven so far: Republish check reported NewsStand SHA-aligned at the snapshot.

Likely

Best guess, not proven: Original fault believed resolved (likely, not verified).

Unknown

Still unknown: Whether the fault recurs across a full observation window.

Known

A worker moves this next; you do not need to do anything.

Case flow · Support ↔ App Builder Hospital
Support homeSupport watchControl Tower

Not flagged escalated. Hub case_escalated evidence appears on Support when observed.

Atlas / Tower return contracts

  • PARTIALatlas/api/public/support-snapshot

    Read-only Support/Pulse snapshot for Atlas /api/atlas/pulse and /api/atlas/support adapters.

    Consumer env: PULSE_SUPPORT_BASE_URL — Payload is code-ready on Pulse. Atlas LIVE-FEEDS-ATLAS-READ-001 wires live fetch when PULSE_SUPPORT_BASE_URL is set; honesty stays PARTIAL until Hub caller + durable store proven. No secrets in snapshot.

  • PARTIALtower/support

    Tower product handoff into Pulse support watch desk (customer-impact observe).

    Consumer env: VITE_SUPPORT_URL — Tower already declares VITE_SUPPORT_URL. Point it at the Pulse origin + /support for watch. Canonical Support is OmniSupport Nexus — Pulse is not a second Support product. Missing URL stays setup-required — not Connected.

Snapshot: GET /api/public/support-snapshot

Outbound handoffs (Pulse → Atlas / Tower)

  • MISSINGOpen Atlas

    Setup required: set VITE_ATLAS_URL to an https URL.

    VITE_ATLAS_URL not set — Atlas return deep-link unavailable.

  • MISSINGOpen Tower

    Setup required: set VITE_TOWER_URL to an https URL.

    VITE_TOWER_URL not set — Tower product return deep-link unavailable. In-Pulse /control-tower still works.

In-Pulse Control Tower observe surface: /control-tower

Build / product repair hop

MISSINGApp Builder Hospital (build/product repair)GAPNOT CONNECTED

HMAC client posts to https://builder.kleinhubai.com/api/public/pulse/health-intake with header x-pulse-signature (sha256=). Shared secret PULSE_INTAKE_SIGNING_SECRET must be set on Pulse and App Builder. Pulse also needs APP_BUILDER_HOSPITAL_WORKSPACE_ID. Live hop remains NOT_CONNECTED until a signed intake is accepted. Pulse watch desks are not App Builder Hospital.

Mission Control / OmniSupport Nexus

MISSINGMission Control / OmniSupport NexusNOT CONNECTEDno GitHub repo

Canonical Support is OmniSupport Nexus (Lovable 530988ff). Emit safe ops signals to Pulse POST /api/public/hooks/omnisupport-signal (Bearer PULSE_SUPPORT_SERVICE_TOKEN). Pulse /support is watch/impact only — not a second ticket DB. Not CONNECTED until live emit proven.

  • Hub support evidence hop · PARTIAL · /api/public/hooks/hub-support-evidence
  • OmniSupport operational signal intake · PARTIAL · /api/public/hooks/omnisupport-signal
  • Tower support URL · PARTIAL · /support
  • OmniSupport Nexus (Lovable) · MISSING
Repair loop
  1. Detect
  2. Triage
  3. Diagnose
  4. Guardian check
  5. Assign specialist
  6. Owner GO
  7. Repair
  8. Test
  9. Audit
  10. Guardian recheck
  11. Recovery watch
  12. Discharge

Pulse snapshot · newsstand · Recent · 2026-08-06 17:30 UTC · Known

Symptoms
  • • Prior warning state
  • • Republish check reported SHA alignment
Confirmed facts
  • • Republish check reported NewsStand SHA-aligned at the snapshot.
Suspected causes (not proven)
  • • Original fault believed resolved (likely, not verified).
Unknowns
  • • Whether the fault recurs across a full observation window.
Why this severity

customer Impact

Readers were affected during the original fault.

data Risk

None observed

security Risk

None observed

revenue Impact

None observed

operational Impact

None observed

release Impact

Recovery build must be verified before discharge.

blast Radius

NewsStand readers

time Sensitivity

Normal

Impact

Affected users: NewsStand readers

Affected systems: NewsStand, University surface

Rollback baseline: Pre-recovery NewsStand release

Verification

Audit verdict: Pending recheck

Guardian recheck: Required before deploy.

Recovery watch: Open — observation window running.

Discharge proof: None recorded

Guardian gate GG-005 Guardian Waiting

Deploy the NewsStand recovery build

Deploying before recovery observation completes could re-introduce the original fault.

Protected area: deployment · Owner decision: Waiting

  • • Recovery watch window completed
  • • Audit verification of the repair
  • • Rollback baseline recorded

Pulse displays this gate. Only the owner can decide it, and only Guardian can clear it.

Repair plan and tests
  1. 1. Observe for the full window
  2. 2. Collect Audit verification
  3. 3. Discharge or reopen
  • NOT RUN
    Post-recovery smokenot run
Work orders
WO-003
Validation
Guardian WaitingNot connected

mightyprophets-rgb/kleinhub-ai-newsstand · main · first command pnpm run test:smoke

Start proof: none — assignment only, no proven start

Timeline
  • 2026-08-06T17:30:00ZRecovery reported; observation opened.

Read-only view. Pulse never publishes, deploys, migrates or clears a Guardian gate.