* feat(evolve): enhance agent reply processing and logging capabilities
- Add agent_reply_result_text function to extract final user-visible text from agent replies
- Implement comprehensive logging throughout auto_memory, auto_resource, and dream modules
- Add max_units configuration option to limit extracted memory units
- Improve error handling and validation in auto_resource step
- Refactor dream extract step to respect max_units limit during processing
- Enhance summary rendering in dream finish step with detailed breakdown
- Add result hook functionality for embedding hosts integration
- Implement loose resource filename handling for root-level resources
- Update test cases to reflect new functionality and improved error messages
* test(background-steps): update fake upsert function to include created parameter
- Modified fake_upsert function to accept 'created' parameter instead of '_created'
- Added 'created' field to captured dictionary in fake_upsert function
- Included 'created': True in the expected response dictionary for test case
- Updated test assertion to match new parameter structure
