AgentExecutorOptions
Options & Configuration · AiDotNet.Agentic.Agents
Settings for an AgentExecutor: identity, system prompt, the tool-loop budget, and the
sampling knobs forwarded to each model call.
For Beginners
These are the dials for one agent. The most useful ones are
SystemPrompt (the agent’s standing instructions / persona) and MaxIterations
(how many think→use-tool→think rounds it may take before it must answer).
How It Works
Every value is nullable and defaults to a sensible behavior when left null, following the
library-wide options pattern (zero-config by default, fully overridable). The executor applies the
documented defaults internally, so the common case is new AgentExecutor<float>(client).
Constructors
| Constructor | Summary |
|---|---|
AgentExecutorOptions() |
Properties
| Property | Summary |
|---|---|
ToolChoiceMode? ToolChoice { get; set; } |
Gets or sets how eagerly the model may use the agent’s tools when tools are present. |
double? Temperature { get; set; } |
Gets or sets the sampling temperature forwarded to each model call. |
int? MaxIterations { get; set; } |
Gets or sets the maximum number of model calls in one run (each tool round costs one call). |
int? MaxOutputTokens { get; set; } |
Gets or sets the per-call output-token cap forwarded to each model call. |
string Description { get; set; } |
Gets or sets the agent’s one-line specialty description. |
string Name { get; set; } |
Gets or sets the agent’s name (used by coordinators and when the agent is surfaced as a tool). |
string SystemPrompt { get; set; } |
Gets or sets the system prompt prepended to every run (the agent’s standing instructions/persona). |
Fields
| Field | Summary |
|---|---|
int DefaultMaxIterations |
The default maximum number of model calls in a single run when MaxIterations is unset. |