Start here: Automate

If you are building agents or automation that mutate or query Asset Core, this path explains how to use the bounded tool surface safely. Asset Core exposes a fixed operation vocabulary and a canonical tool catalog so automated systems can act without risking uncontrolled mutations.

This path is designed to give you the operational guardrails first, then the specific tooling details and schemas. It is the fastest route to a safe, auditable agent integration.

Recommended path

1. Agents Overview - Why bounded operations are the safe surface.
2. Tooling Guide - Canonical tool catalog across MCP/OpenAI/Gemini.
3. Action Reference - Full operation reference for tooling.
4. HTTP API - Underlying endpoints each tool uses.
5. Transactions and Operations - Mental model for commits.

What success looks like

A correct automation integration can submit a transaction, read back freshness-aware state, and interpret errors without leaking unsafe actions. You should also be able to audit every tool call via the commit history endpoints.

Next steps

• If you are running Asset Core in production, switch to the Operate path.
• If you are integrating an application directly, switch to Start here: Integration.