Struct FoundryExecutor 

pub struct FoundryExecutor;

Foundry/Azure OpenAI executor implementing the Executor trait.

Supports two connection kinds:

• ApiKey: Uses the api-key header for authentication
• Foundry: Uses the Foundry endpoint (Entra ID auth, when available)

Falls back to API key auth when no explicit connection is configured.

Trait Implementations

impl Executor for FoundryExecutor

fn execute<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, agent: &'life1 Prompty, messages: &'life2 [Message], ) -> Pin<Box<dyn Future<Output = Result<Value, InvokerError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Execute an LLM call with the given messages.

fn format_tool_messages( &self, _raw_response: &Value, tool_calls: &[ToolCall], tool_results: &[String], _text_content: Option<&str>, ) -> Vec<Message>

Format tool-call results into messages for the next iteration of the agent loop. Returns messages to append to the conversation.

fn execute_stream<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, agent: &'life1 Prompty, messages: &'life2 [Message], ) -> Pin<Box<dyn Future<Output = Result<Pin<Box<dyn Stream<Item = Value> + Send>>, InvokerError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Execute an LLM call and return a stream of raw SSE chunks.