releaseintermediate
[Release] crewaiInc/crewAI 1.14.4a1: 1.14.4a1
By greysonlalondegithub
View original on githubcrewAI v1.14.4a1 is a pre-release update focusing on bug fixes and documentation improvements. Key fixes address LLM chat failures, message state management in executors, trained-agent file handling, multimodal input support, and JSON serialization issues. The release includes new documentation for sandbox tools (E2B, Daytona), Vertex AI setup, and You.com MCP integration, plus security updates to litellm.
Key Points
- •Fixed crew chat description helpers to handle LLM failures gracefully
- •Reset messages and iterations between executor invocations to prevent state carryover
- •Ensure trained-agents files are properly forwarded through replay, test, and inference workflows
- •Bind task-only agents to crew to enable multimodal input_files to reach the LLM
- •Serialize guardrail callables as null for proper JSON checkpointing compatibility
- •Renamed `force_final_answer` parameter in agent_executor to eliminate self-referential routing issues
- •Updated litellm dependency for SSTI security fix and CVE management
- •Added comprehensive documentation for E2B and Daytona sandbox tools integration
- •Documented Vertex AI workload identity setup for cloud deployments
- •Integrated You.com MCP tools for search, research, and content extraction capabilities
Found this useful? Add it to a playbook for a step-by-step implementation guide.
Workflow Diagram
Start Process
Step A
Step B
Step C
Complete