Agent SDK
SDK primitives for building bounded financial agents on top of Ryvra policies, identities, and deterministic command surfaces.
Agent SDK packages identity binding, mandate-aware requests, simulation hooks, and provenance capture so teams can integrate quickly without inventing their own unsafe agent control model.
What it enables
Typed command surfaces
Use SDK helpers for payments, markets, accounts, and treasury actions with clear lifecycle expectations.
Policy-aware client flows
Submit requests with identity, mandate, idempotency, and policy metadata wired in from the start.
Provenance capture
Attach correlation and evidence fields that simplify audit and incident response.
Integration acceleration
Reduce custom glue code across agent products and enterprise finance systems.
Control model
Constraints this module holds regardless of what an agent proposes.
- SDK flows are designed for bounded autonomy, not unrestricted trading or wallet control.
- Auth and signing remain operator-managed and rotation-friendly.
- Commands stay subject to gateway, policy, risk, and settlement checks.
- Observability hooks expose request, event, and settlement provenance.
Where it fits
Developers
Ship agentic finance integrations faster with standard authority and evidence patterns.
Enterprise platform teams
Standardize safe agent integrations across internal business units and products.