Module agent 

Structs

AgentCheckpointingConfig

AgentCodexAppServerConfig

AgentConfig

Agent-wide configuration

AgentHarnessConfig

AgentOnboardingConfig

AgentPromptSuggestionsConfig

Inline prompt suggestion configuration for the chat composer.

AgentSmallModelConfig

Small/lightweight model configuration for efficient operations

AgentVibeCodingConfig

Vibe coding configuration for lazy/vague request support

AsyncApprovalConfig

Configuration for async (out-of-band) approval of tool execution requests.

CircuitBreakerConfig

ConfidenceEscalationConfig

Implements the escalation decision rule from “The Hitchhiker’s Guide to Agentic AI” (Eq. 18.12): Escalate iff p_success < tau_conf OR action in A_irreversible OR cost > B_auto

MemoriesConfig

Codex-compatible memories configuration.

OpenResponsesConfig

Open Responses specification compliance configuration

PersistentMemoryConfig

SkepticPanelConfig

Configuration for the adversarial multi-model evaluator panel.

ToolResultClearingConfig

TrackerContinuationConfig

Tracker-aware auto-continuation policy.

VerificationAutoRecoveryConfig

Autonomous recovery policy for the anti-blind-editing verification gate.

Enums

ContextResetMode

When to trigger a context reset — starting a clean session from external artifacts only, discarding conversation history to clear noise and bad assumptions. This is distinct from compaction, which preserves conversational continuity within the same task/agent loop.

ContinuationPolicy

HarnessOrchestrationMode