Overview
AgentKit applies retry logic during agent execution, but the default rules are not clearly documented. Developers may need to inspect implementation details to understand when and how retries occur.
Problem Description
Without clear guidance, it is difficult to know which failures trigger retries, how many retry attempts are made, or how retries affect execution flow and timing.
Expected Behavior
The documentation should clearly describe the conditions under which retries are performed and how retry limits or backoff are applied.
Proposed Solution
Add a short documentation section that explains the default retry rules, including which errors are retryable, how many attempts are made, and how retries impact agent execution.