II.
AgentCoreImpl overview
Reference · liveagent-core-impl:claude-agent-sdk.core@current
Claude Agent SDK Core (current) overview
Inspect the raw attributes, linked wiki pages, and inbound or outbound graph edges for agent-core-impl:claude-agent-sdk.core@current.
Attributes
displayName
Claude Agent SDK Core (current)
agentVersionId
packageRef
loopIteratorPolicy
process-function-driven
loopIteratorNotes
Host program imports `@anthropic-ai/claude-agent-sdk` and consumes
the `query()` async iterator. Each iteration yields SDKMessage
events; the host decides when to feed back permission/elicitation
results. There is no embedded chat-loop binary — the loop body
IS the host's code.
contextManagementStrategy
server-managed
compactionTriggerNotes
Compaction is handled by the Claude server-side; the SDK does not
expose a token-threshold trigger.
subagentInvokerPolicy
platform-mediated
subagentInvokerNotes
Subagents (when enabled) are dispatched by the SDK via the same
Task-tool primitive as Claude Code; the host program mediates
permission gating.
resultSynthesisPolicy
model-direct
resultSynthesisNotes
Final result events arrive as SDKMessage values on the iterator;
the host program emits the final response.
stopDetectionStrategy
structured-end-event
transportClientLibrary
@anthropic-ai/claude-agent-sdk
supportedTransportProtocols
parallelToolCallHandling
native
streamingFidelity
full
thinkingChannelHandling
passthrough
loopBudgetKnobs
- max_turns
- max_budget_usd
compactionTrigger
auto-on-budget
streamingBlocks
- text
- thinking
- tool_use
effortLevelDelegation
native
Outgoing edges
composes1
- agent:claude-agent-sdk·AgentProductClaude Agent SDK
emits_message_type9
- protocol-message:system-message·ProtocolMessageSDKMessage / SystemMessage
- protocol-message:assistant-message·ProtocolMessageSDKMessage / AssistantMessage
- protocol-message:user-message·ProtocolMessageSDKMessage / UserMessage
- protocol-message:stream-event·ProtocolMessageSDKMessage / StreamEvent
- protocol-message:result-message-success·ProtocolMessageResultMessage subtype=success
- protocol-message:result-message-error-max-turns·ProtocolMessageResultMessage subtype=error_max_turns
- protocol-message:result-message-error-max-budget-usd·ProtocolMessageResultMessage subtype=error_max_budget_usd
- protocol-message:result-message-error-during-execution·ProtocolMessageResultMessage subtype=error_during_execution
- protocol-message:result-message-error-max-structured-output-retries·ProtocolMessageResultMessage subtype=error_max_structured_output_retries
fires_hook6
- hook-surface:pre-tool-use·HookSurfacePreToolUse (canonical)
- hook-surface:post-tool-use·HookSurfacePostToolUse (canonical)
- hook-surface:user-prompt-submit·HookSurfaceUserPromptSubmit (canonical)
- hook-surface:stop·HookSurfaceStop (canonical)
- hook-surface:subagent-stop·HookSurfaceSubagentStop (canonical)
- hook-surface:pre-compact·HookSurfacePreCompact (canonical)
realizes1
- layer:4-agent-core·LayerAgent-Core
speaks1
- model-transport:anthropic-messages·ModelTransportProtocolAnthropic Messages
supports3
- capability:streaming·CapabilityStreaming
- capability:supports-tool-use·CapabilitySupports tool use
- capability:supports-thinking·CapabilitySupports thinking
supports_effort_level5
- effort-level:low·EffortLevelEffort: low
- effort-level:medium·EffortLevelEffort: medium
- effort-level:high·EffortLevelEffort: high
- effort-level:xhigh·EffortLevelEffort: xhigh
- effort-level:max·EffortLevelEffort: max
supports_orchestration_primitive3
- orch-primitive:agent-dispatch·OrchestrationPrimitiveAgent Dispatch
- orch-primitive:budget-stop·OrchestrationPrimitiveBudget Stop
- orch-primitive:cost-tracking·OrchestrationPrimitiveCost Tracking
Incoming edges
composed_of1
- agent-version:claude-agent-sdk@current·AgentVersion