Skip to main content

​
HandoffConfig

Defined in the handoff module.
AI Agent Unified configuration for handoff behavior. This consolidates all handoff-related settings including context policy, timeouts, concurrency control, and safety features. Attributes: context_policy: How to share context during handoff (default: summary for safety) max_context_tokens: Maximum tokens to include in context max_context_messages: Maximum messages to include (for LAST_N policy) preserve_system: Whether to preserve system messages in context tool_policy: Tool boundary enforcement policy (NEW - security by default) timeout_seconds: Timeout for handoff execution max_concurrent: Maximum concurrent handoffs (0 = unlimited) detect_cycles: Enable cycle detection to prevent infinite loops max_depth: Maximum handoff chain depth async_mode: Enable async execution allow_parallel: Enable parallel execution (from Delegator) on_handoff: Callback when handoff starts on_complete: Callback when handoff completes on_error: Callback when handoff fails

​
Properties

​
ContextPolicy
No description available.
​
int
No description available.
​
int
No description available.
​
bool
No description available.
​
HandoffToolPolicy
No description available.
​
float
No description available.
​
int
No description available.
​
bool
No description available.
​
int
No description available.
​
bool
No description available.
​
bool
No description available.
​
Optional[Callable]
No description available.
​
Optional[Callable]
No description available.
​
Optional[Callable]
No description available.
  • from_dict: Create config from dictionary.
  • to_dict: Convert config to dictionary.

​
Source

View on GitHub

praisonaiagents/agent/handoff.py at line 70

Configuration Overview

Agent Config

Handoffs Concept

Handoffs Feature