Getting Started (Users and Operators)

This guide helps you get from first visit to first controlled agent workflow in Operus.

What this guide is for

This guide explains how to interact with current Operus product surfaces safely.

It does not assume that every workflow is fully autonomous, fully public, or fully live today. Always interpret actions and metrics according to their documented trust stage.

Before you begin

• Use a supported wallet setup.
• Confirm you are connected to the expected network configuration.
• Understand that product surfaces can include both onchain-real and simulation-backed data, depending on context.

Step 1: Connect wallet

1. Open Operus and connect your wallet.
2. Verify account and network details.
3. Ensure your operator/admin permissions match the actions you intend to perform.

Step 2: Create your first agent

1. Open the create-agent flow.
2. Set agent metadata and strategy context.
3. Save and review the generated agent profile.

At this stage, treat strategy setup as configuration, not proof of live strategy quality.

Step 3: Review agent page semantics

On each agent page, always check:

• mode context (simulation, paper, live-staged)
• confidence and evidence signals
• proposal/decision status
• freshness context where quote/build data is shown

Step 4: Use operator controls safely

For any sensitive action:

1. Review proposal details and risk context.
2. Confirm routing/quote freshness and policy constraints.
3. Approve, reject, or hold with explicit operator intent.
4. Verify lifecycle events and audit records were written.

If any evidence is missing or stale, prefer hold over approve.

Step 5: Monitor lifecycle and audit history

After decisions, check:

• proposal status transitions
• decision event history
• lifecycle timeline facts
• reconciliation fields (where available)

This is your source of truth for operational replay.

Common first-week checklist

• Create one agent and validate metadata flow.
• Trigger one non-sensitive proposal flow and inspect evidence output.
• Review one operator decision path end-to-end.
• Confirm you can distinguish simulation-backed and trust-staged data.

Safety defaults

• Start small and controlled.
• Keep approvals explicit.
• Prefer verifiable traces over speed.
• Escalate only after stable behavior and clean lifecycle evidence.

What not to assume

• Do not assume all displayed performance is live-verified.
• Do not assume all sensitive actions execute automatically.
• Do not assume strategy setup is proof of strategy quality.
• Do not assume every approval implies autonomous execution.

Use labels, evidence, and lifecycle history as the source of truth.