Module context

Expand description

Core traits and types.

Always available regardless of feature flags. Includes:

Agent - The fundamental trait for all agents
Tool / Toolset - For extending agents with capabilities
Session / State - For managing conversation context
Event - For streaming agent responses
AdkError / Result - Unified error handling Invocation context traits: state, session, artifacts, memory, and run configuration.

Structs§

MemoryEntry: A single entry returned from memory search.
RunConfig: Configuration for a single agent run.
RunConfigBuilder: Builder for RunConfig.
ToolCallbackContext: Wraps a CallbackContext to inject tool name and input for before-tool and after-tool callbacks.
ToolConcurrencyConfig: Configuration for tool execution concurrency.
ToolConfirmationRequest: Payload describing a tool call awaiting human confirmation.
ToolOutcome: Structured metadata about a completed tool execution.

BackpressurePolicy: Policy for handling excess tool calls when the concurrency limit is reached.
IncludeContents: Controls what parts of prior conversation history is received by llmagent
StreamingMode: Streaming mode for agent responses. Matches ADK Python/Go specification.
ToolConfirmationDecision: Decision applied when a tool execution requires human confirmation.
ToolConfirmationPolicy: Policy defining which tools require human confirmation before execution.

Artifacts: Binary artifact storage for agents.
CallbackContext: Context available to agent lifecycle callbacks.
InvocationContext: Full invocation context available to agents during execution.
Memory: Semantic memory search for agents.
ReadonlyContext: Read-only access to invocation metadata.
ReadonlyState: Read-only view of session state.
SecretService: Trait for retrieving secrets at runtime.
Session: Represents an active conversation session with identity and state.
State: Mutable session state with key-value storage.

validate_state_key: Validates a state key. Returns Ok(()) if the key is safe, or an error message.