Agent Use Pack v1

Agents use Wever Labs rails from one start surface.

This is the canonical entry path for agents and builders: discover PacketOps and DiligenceOps, quote a run, receive a payment challenge, start the rail, check status, fetch the return package, and record the receipt.

Open Agent Use API Schema Start-run example Sample packages

DiscoverQuotePayment challengeStart runStatusReturn packageReceipt

Available now

One rail call, one proof path.

Use /api/agent-use for the full machine-readable route. Public read calls stay open. Sandbox writes return usable records without a key. Production writes use Production Write Gate scoped keys, idempotency keys, and audit events when persistence is enabled.

GET /api/agent-use?compact=true
POST /api/agent-use
{
  "schema_version":"agent_use.v1",
  "mode":"start_agent_run",
  "rail_key":"packetops",
  "idempotency_key":"agent-run-001",
  "payment_reference":"checkout_or_wallet_ref"
}

First supported rails

Start with bounded operational work.

Agents do not need to decode the site. They choose a rail, pass the required context, attach payment or payment reference, and receive a package target.

PacketOps

Packet readiness

Best for document packets, onboarding sets, submission folders, and missing-item checks.

Returns: manifest, missing-item report, readiness object, exception object, attestation.

Read PacketOps route

DiligenceOps

Evidence review

Best for vendor evidence, diligence folders, readiness scoring, and review package preparation.

Returns: evidence inventory, missing-item report, risk summary, readiness score, attestation.

Read DiligenceOps route

Machine call sequence

Use the rail in seven small moves.

The sequence is deliberately boring. Boring is the compliment. Agents need exact routes, stable fields, status URLs, and receipts.

Read start pack

GET /api/agent-use returns rails, modes, pricing references, and schemas.

Quote run

mode=quote_run returns amount, currency, package level, and accepted payment methods.

Payment challenge

mode=create_payment_challenge returns a 402-style JSON envelope with amount, run ID, expiry, and accepted methods.

Start run

mode=start_agent_run requires rail key, idempotency key, and a payment or wallet reference.

Check status

mode=get_run_status or GET /api/agent-use?run_id=... returns the current run state.

Fetch package

mode=fetch_return_package returns the package contract and proof fields.

Record receipt

mode=record_agent_use_receipt binds receipt ID, payment reference, return package ID, and hash.

Payment challenge

Payment and work move as one bounded object.

The payment challenge is intentionally portable. It can point to hosted checkout, wallet allowance, invoice reference, stablecoin reference, x402 reference, or AP2-style mandate reference.

{
  "payment_required": true,
  "challenge_id": "payment_challenge_...",
  "rail_key": "packetops",
  "amount_units": 99,
  "currency": "usd",
  "accepted_methods": ["hosted_checkout", "agent_allowance", "payment_reference", "invoice_reference", "stablecoin_reference", "x402_reference", "ap2_mandate_reference"],
  "status_url": "/api/agent-use?run_id=...",
  "receipt_required": true
}

Use the rail

Agents can start with PacketOps or DiligenceOps now.

For production runs, attach a payment or wallet reference. For sandbox runs, send the same shape and keep the receipt trail intact.

Open API Open Live Sprint Production Write Gate