The agent builder tabs
What each tab does, from Basic info and Instructions through Guardrails, Approval, Schema, Runtime and Publish.
Basic info. Name, description and the agent's role inside apps.
Instructions. The system prompt. It is sent with every model call unless a workflow node overrides it. Keep it specific: audience, tone, what to do when information is missing, when to use tools.
Model. Provider, model, temperature, max output tokens, max iterations (how many model→tool→model loops a run may take). See Models and providers.
Tools. Attach tools from the organisation library. Attached tools are inherited by the agent's toolbox node. Tool settings such as authentication, safety level and limits live on the tool itself (Tools).
Knowledge. Attach knowledge bases. The knowledge retrieval node uses them by default.
Memory. Memory mode and defaults: off, remember in this chat, remember for this agent, remember this user; plus the namespace and read/write defaults (Memory).
Workflow. The graph editor: nodes, edges, conditional edges, per-node settings, validation, and versions (Workflows: nodes, edges and the default graph).
Logs. Runs of this agent with every step, prompt, tool call and retrieval (Testing, logs and observability).
Guardrails. Blocked topics and deny/allow rules checked before the final answer (Guardrails, approvals and safety).
Approval. Which tools require approval before they execute (Guardrails, approvals and safety).
Schema. An optional output schema. When set, the workflow validates the answer as JSON against the schema and can retry or repair it (validation node, Workflows: nodes, edges and the default graph).
Runtime. Runtime settings for the agent: retry policy, timeouts and streaming behaviour.
Publish. Publish the draft, see published versions, restore an earlier version.