The done_output field has max_length=100000, but agent.history.final_result()
can return arbitrarily long strings (e.g., large data extractions). This caused
Pydantic validation errors when constructing UpdateAgentTaskEvent at the end of
agent.run(), crashing the entire run even though the task itself succeeded.
Truncate done_output to MAX_STRING_LENGTH before passing to the constructor.
Co-authored-by: Magnus Müller <MagMueller@users.noreply.github.com>
The ARN of cross-region inference exceeds the limit of 100 characters, which is not allowing to make ChatAnthropicBedrock calls. The limit increase ensures we are able to use the inference endpoints with the ARN.
- Updated `run_auth_command` to provide detailed authentication status updates and debug information during the authentication process.
- Introduced a new `UpdateAgentSessionEvent` class to facilitate session updates.
- Enhanced `CloudSync` to allow all events during the authentication flow and to show direct links for authenticated users.
- Improved error handling and user messaging for authentication failures.
- Updated documentation to clarify cloud sync settings and authentication instructions.
Magnus Müller