Zenith Hypernova · AGI Jobs

Authorized work. Reviewable evidence. Accountable scale.

AGI Jobs is designed as a scalable machine labor layer for authorized, lawful screen-based work—coordinating specialized agents to execute tasks, produce reviewable evidence, and support independent verification and settlement across a broad range of computer-based workflows.

Open the interactive workbench · Runbook · Owner controls · Execution integration

Start in two minutes

The browser workbench needs no wallet, credentials or installation. Select Analyze the mission, inspect the calculations, then export the evidence, CSV and readable report. Inspect the original defects compares the preserved historical input. Create a brief exports one of ten computer-work proposals with explicit acceptance tests and an unfunded USDC reward ceiling. Keep Download task source (source-plan.json) with each work order and operator brief; verify its SHA-256 before execution. Browser proposals use the corrected mission, while analysis imports have a separate source export. The CLI supports custom proposal sources with --source.

For local operation, use the repository .nvmrc (Node 22.23.3):

npm run demo:zenith-hypernova:serve
# Open http://127.0.0.1:4178
npm run demo:zenith-hypernova:work
npm run demo:zenith-hypernova:task -- --type governance-audit --region EARTH
npm run demo:zenith-hypernova:test

The new analysis, server and unit tests use only Node built-ins. The preserved contract kit and browser QA require the root locked dependencies (npm ci). Run from the repository root. CLI output goes to a new directory under reports/zenith-hypernova-work/; --out selects a new destination and refuses to overwrite prior evidence.

What is implemented

Path	Actual behavior	Boundary
Browser workbench / `:work`	Computes exact budgets and a dependency schedule; exports JSON, CSV, Markdown and byte-bound evidence	Synthetic plan analysis; no provider or chain calls
Work-order builder / `:task`	Ten templates, six research contexts, acceptance criteria, source hash and integer USDC ceiling	Proposal only; owner must approve sources, runtime, spending and reviewer
Evidence checker / `:review`	Recomputes arithmetic and scheduling using a separate implementation, checks every output and source hash	Content checks do not establish reviewer independence or buyer acceptance
Capacity model	Limits admitted work by demand, workers and reviewer hours; applies an acceptance-rate assumption	Illustrative throughput; excludes costs, rework, disputes and downtime
`demo:zenith-hypernova`	Preserved ASI Global governance-kit generator with Hypernova plan and report scope	Shared protocol rehearsal, not execution of the plan’s 11 physical projects
`demo:zenith-hypernova:local`	Isolated localhost production-contract lifecycle using the shared ASI Global mission specification	Local mock assets and receipts; not mainnet or paid customer work

The $40T/year ceiling is a user-supplied scenario assumption, not an independently established market estimate, revenue forecast or guarantee of reachable demand. Scaling toward planetary and stellar infrastructure is a long-term ambition; the implemented entry point is useful, verifiable digital work.

Corrected scenario, preserved history

All six regions, 11 jobs, dependencies, rewards, durations, participant roles and original Mermaid source remain. The original input is available byte-for-byte in fixtures/legacy-project-plan.json.

The original plan allocated 1.173 billion of 1.25 billion scenario units. Regional rewards exceeded allocations by 7 million (Africa), 28 million (MENA) and 26 million (Earth). The correction assigns 61 million of the 77 million unallocated units to those regions and adds the remaining 16 million to the governance reserve. Total planned rewards remain 1.128 billion; remaining budget is 122 million. No job reward was changed.

deadlineDays now explicitly means stage duration after all dependencies are accepted. Under unlimited parallel resources, the critical path is 286 days. It is not an absolute due date or a real infrastructure-delivery prediction. AGIALPHA is retained as the legacy synthetic accounting unit. New work proposals use USDC; this demo does not implement or certify a USDC settlement contract.

Systems map — original source preserved

Architecture diagram · source preserved below

View original Mermaid source

flowchart LR
    Operators((Mission Owners)) --> demo_zenith_sapience_initiative_supra_sovereign_hypernova_governance[[Demo → Zenith Sapience Initiative Supra Sovereign Hypernova Governance]]
    demo_zenith_sapience_initiative_supra_sovereign_hypernova_governance --> Core[[AGI Jobs v0 (v2) Core Intelligence]]
    Core --> Observability[[Unified CI / CD & Observability]]
    Core --> Governance[[Owner Control Plane]]

From work to accountable scale

Architecture diagram · source preserved below

View original Mermaid source

flowchart TD
    Scope["Approved scope and inputs"] --> Creator["Specialist execution"]
    Controls["Owner limits and stop controls"] --> Creator
    Creator --> Evidence["Artifacts, sources and tests"]
    Evidence --> Reviewer["Independent review"]
    Reviewer -->|"Corrections"| Scope
    Reviewer --> Buyer["Buyer acceptance"]
    Buyer --> Settlement["Authorized settlement"]
    Controls --> Settlement
    Settlement --> Capacity["Reinvest verified gains"]
    Capacity --> Scope

Useful software, research, analysis and editable documents are the first layer. Reliable delivery can support energy and industrial research; proven improvements can finance more compute and productive capacity. Each stage needs its own evidence and authority.

Verification and governance

Every change lands through a pull request. Required checks must pass before merge. The Hypernova workflow exercises the retained kit/local-chain paths and the new analysis tests; Pages checks the published browser experience, mobile layouts and accessibility. The validation record states the executed scope and remaining live commissioning requirements.

Use RUNBOOK.md for exact commands, OWNER-CONTROL.md for preview and signing boundaries, and INTEGRATION.md for current OpenClaw / OpenAI / ChatGPT Work handoff guidance. Secrets stay outside the repository and public evidence. Original governance kits, diagrams and local receipts remain inspectable.