# Sharvora Agent Workflows

All workflows are advisory. `automatic_purchase` is always `false`.

## Agent Trust
Assess an unfamiliar endpoint before an autonomous agent trusts it or spends against it.

1. `GET /api/v1/router/recommend?goal=assess%20an%20unfamiliar%20agent%20endpoint` — Choose the minimum useful trust check.
2. `POST /api/v1/web/agent-readiness` — Check whether the target exposes machine-readable agent/API discovery.
3. `POST /api/v1/web/endpoint-trust` — Inspect endpoint behavior for redirect, prompt-injection, and credential-exfiltration risk.
4. `POST /api/v1/web/trust-passport` — Escalate only when a combined trust package is justified.

Machine-readable: https://sharvora.com/api/v1/workflows

## Decision / Research
Preflight a decision, synthesize multiple quantitative signals, and later audit the outcome.

1. `POST /api/v1/quote` — Validate the selected product and body structure without spending.
2. `POST /api/v1/decision/preflight` — Check freshness and missing context before acting.
3. `POST /api/v1/decision/bundle` — Combine risk, anomaly, forecast, and decision evidence in one paid call.
4. `POST /api/v1/decision/audit` — Evaluate the recorded decision after an outcome is observable.

Machine-readable: https://sharvora.com/api/v1/workflows

## Validate Before Spend
Inspect request shape and x402 payment metadata before an agent authorizes a purchase.

1. `GET /api/v1/request-guide` — Read the exact method, schema, and example body.
2. `POST /api/v1/quote` — Validate structure and price without creating a payment.
3. `POST /api/v1/x402/validate` — Inspect a third-party x402 challenge when protocol validation is worth the $0.005 check.
4. `GET /api/v1/service-proof` — Use the low-cost Sharvora canary only when an end-to-end paid delivery proof is needed.

Machine-readable: https://sharvora.com/api/v1/workflows