Superintelligence Empowerment Playbook

The Planetary Orchestrator Fabric demo is designed so a non-technical owner can direct a superintelligent, planetary-scale workforce without writing code. This playbook distills the mission into decisive actions, verifications, and recovery drills that prove AGI Jobs v0 (v2) grants total operational dominance.

Mission Timeline

Architecture diagram · source preserved below

View original Mermaid source

timeline
  title Planetary Orchestrator Fabric – Operator Timeline
  section Launch
    Calibrate fabric config : Load `config/fabric.example.json`
    Ignite planetary run : `bin/run-demo.sh --jobs 10000`
    Inspect live dashboard : Open `reports/<label>/dashboard.html`
  section Governance
    Execute owner schedule : `config/owner-commands.example.json`
    Reroute & throttle shards : `owner-script.json`
    Tighten checkpoint cadence : `checkpoint.configure`
  section Resilience
    Trigger outage drill : `--simulate-outage mars.gpu-helion`
    Halt after X ticks : `--stop-after-ticks`
    Resume from snapshot : `--resume --checkpoint <path>`
  section Assurance
    Run acceptance autopilot : `npm run demo:planetary-orchestrator-fabric:acceptance`
    Inspect ledger invariants : `reports/<label>/ledger.json`
    Publish mission dossier : Archive `reports/<label>` bundle

Operator Checklist

Prime the environment using the quickstart in the README. Use the repository-pinned Node/npm versions and run npm ci once. The simulator and local workbench need no credentials.
Launch the fabric with bin/run-demo.sh. Supply --output-label to route artifacts into a named directory.
Load the mission console (ui/dashboard.html) and drop the generated reports/<label> folder onto the page to review topology, job flows, owner interventions, and restart drill guidance without a build step.
Replay governance payloads from reports/<label>/owner-script.json or generate new ones with npm run owner:system-pause etc.
Rehearse the restart drill via bin/run-restart-drill.sh --stop-after 200. The script halts the orchestrator, extracts the checkpoint path from summary.json, and resumes automatically.
Validate planetary resilience by executing npm run demo:planetary-orchestrator-fabric:acceptance -- --label planetary-acceptance. The harness proves that:
- <2% of jobs drop during 10k-load with a simulated node outage.
- The orchestrator can be killed mid-run and resume from checkpoint without duplication or loss.
File the audit bundle by shipping summary.json, events.ndjson, ledger.json, mission-topology.mmd/mission-topology.html, and the owner command logs to your archival system.

Multi-Angle Verification Grid

Perspective	Command	Evidence Produced
Deterministic execution	`npm run test:planetary-orchestrator-fabric`	Validates shard balance, node outage recovery, deterministic replay, and owner command supremacy.
CI parity	GitHub workflow `demo-planetary-orchestrator-fabric`	Enforces lint, tests, CI-mode demo, acceptance autopilot, and artifact validation on every PR and on `main`.
Production rehearsal	`bin/run-restart-drill.sh`	Demonstrates controlled crash/restart with merged telemetry for auditors.
Governance resilience	`config/owner-commands.example.json` (executed via CLI)	Exercises pause/resume, shard updates, node registration, checkpoint rotation, and reporting retargeting mid-run.
Ledger integrity	Inspect `reports/<label>/ledger.json`	Shows totals, spillover counts, reassignments, and invariants that must all pass.

Owner Control Surface Quick Reference

These are examples of the wider repository owner surface, not authenticated controls exposed by this simulator. Use the JSON command schedule for local rehearsals.

Pause / resume everything: owner:system-pause, owner:mission-control --action resume.
Shard tuning: owner:command-center --action shard-update --shard <id> --queue-alert <value>.
Node lifecycle: owner:command-center --action node-register --node-file <json> or --action node-deregister --node-id <id>.
Job surgery: Use job.reroute or job.cancel locators to move or remove work from any shard, including in-flight tasks.
Checkpoint mastery: Combine checkpoint.configure and checkpoint.save commands to rotate storage paths and snapshot cadence instantly.
Reporting retarget: reporting.configure ensures dashboards, summaries, and event logs follow the owner’s preferred archive without downtime.

Pitfall Radar & Countermeasures

Massive backlog growth: Configure spillover policies in config/fabric.example.json and confirm they activate via events.ndjson spillover entries.
Node churn: Watch getHealthReport() output or the dashboard’s node panel to confirm nodes that miss heartbeats are quarantined and their jobs reassigned.
Checkpoint drift: After checkpoint.configure, run checkpoint.save and verify the new path in summary.json.ownerState.checkpoint.
Report hygiene: Ensure summary.json, events.ndjson, owner-commands-executed.json, and ledger.json live under the retargeted directory—CI will fail if any artifact is missing.
Shard imbalance: Consult reports/<label>/summary.json.shards.*.queueDepth and execute job.reroute to rebalance within milliseconds.

Final Reflective Pass

Walk the entire flow once more—from configuration through restart drill—and confirm every artifact updates as expected. If any invariant or log looks suspicious, re-run the acceptance autopilot to reproduce deterministic telemetry. The resulting evidence qualifies the rehearsed simulator scenarios. Mainnet deployment still requires authenticated control, commissioned workers and independent reviewers, remote policy/stop enforcement, and validated contract settlement. See the computer-work runbook.