feat: use force_final_answer prompt on timeout

Co-Authored-By: Joe Moura <joao@crewai.com>

README.md

@@ -1,10 +1,18 @@
 <div align="center">
 
-![Logo of CrewAI, two people rowing on a boat](./docs/crewai_logo.png)
+![Logo of CrewAI](./docs/crewai_logo.png)
 
 # **CrewAI**
 
-🤖 **CrewAI**: Production-grade framework for orchestrating sophisticated AI agent systems. From simple automations to complex real-world applications, CrewAI provides precise control and deep customization. By fostering collaborative intelligence through flexible, production-ready architecture, CrewAI empowers agents to work together seamlessly, tackling complex business challenges with predictable, consistent results.
+**CrewAI**: Production-grade framework for orchestrating sophisticated AI agent systems. From simple automations to complex real-world applications, CrewAI provides precise control and deep customization. By fostering collaborative intelligence through flexible, production-ready architecture, CrewAI empowers agents to work together seamlessly, tackling complex business challenges with predictable, consistent results.
+
+**CrewAI Enterprise**
+Want to plan, build (+ no code), deploy, monitor and interare your agents: [CrewAI Enterprise](https://www.crewai.com/enterprise). Designed for complex, real-world applications, our enterprise solution offers:
+
+- **Seamless Integrations**
+- **Scalable & Secure Deployment**
+- **Actionable Insights**
+- **24/7 Support**
 
 <h3>
 
@@ -392,7 +400,7 @@ class AdvancedAnalysisFlow(Flow[MarketState]):
             goal="Gather and validate supporting market data",
             backstory="You excel at finding and correlating multiple data sources"
         )
-        
+
         analysis_task = Task(
             description="Analyze {sector} sector data for the past {timeframe}",
             expected_output="Detailed market analysis with confidence score",
@@ -403,7 +411,7 @@ class AdvancedAnalysisFlow(Flow[MarketState]):
             expected_output="Corroborating evidence and potential contradictions",
             agent=researcher
         )
-        
+
         # Demonstrate crew autonomy
         analysis_crew = Crew(
             agents=[analyst, researcher],

docs/concepts/knowledge.mdx

@@ -91,7 +91,7 @@ result = crew.kickoff(inputs={"question": "What city does John live in and how o
 ```
 
 
-Here's another example with the `CrewDoclingSource`. The CrewDoclingSource is actually quite versatile and can handle multiple file formats including TXT, PDF, DOCX, HTML, and more. 
+Here's another example with the `CrewDoclingSource`. The CrewDoclingSource is actually quite versatile and can handle multiple file formats including MD, PDF, DOCX, HTML, and more. 
 
 <Note>
   You need to install `docling` for the following example to work: `uv add docling`
@@ -152,10 +152,10 @@ Here are examples of how to use different types of knowledge sources:
 
 ### Text File Knowledge Source
 ```python
-from crewai.knowledge.source.crew_docling_source import CrewDoclingSource
+from crewai.knowledge.source.text_file_knowledge_source import TextFileKnowledgeSource
 
 # Create a text file knowledge source
-text_source = CrewDoclingSource(
+text_source = TextFileKnowledgeSource(
     file_paths=["document.txt", "another.txt"]
 )
 

docs/concepts/llms.mdx

@@ -463,26 +463,32 @@ Learn how to get the most out of your LLM configuration:
 
   <Accordion title="Google">
     ```python Code
-    # Option 1. Gemini accessed with an API key.
+    # Option 1: Gemini accessed with an API key.
     # https://ai.google.dev/gemini-api/docs/api-key
     GEMINI_API_KEY=<your-api-key>
 
-    # Option 2. Vertex AI IAM credentials for Gemini, Anthropic, and anything in the Model Garden.
+    # Option 2: Vertex AI IAM credentials for Gemini, Anthropic, and Model Garden.
     # https://cloud.google.com/vertex-ai/generative-ai/docs/overview
     ```
 
-    ## GET CREDENTIALS 
+    Get credentials:
+    ```python Code
+    import json
+
     file_path = 'path/to/vertex_ai_service_account.json'
 
     # Load the JSON file
     with open(file_path, 'r') as file:
         vertex_credentials = json.load(file)
 
-    # Convert to JSON string
+    # Convert the credentials to a JSON string
     vertex_credentials_json = json.dumps(vertex_credentials)
+    ```
 
     Example usage:
     ```python Code
+    from crewai import LLM
+
     llm = LLM(
         model="gemini/gemini-1.5-pro-latest",
         temperature=0.7,

docs/concepts/memory.mdx

@@ -58,41 +58,107 @@ my_crew = Crew(
 ### Example: Use Custom Memory Instances e.g FAISS as the VectorDB
 
 ```python Code
-from crewai import Crew, Agent, Task, Process
+from crewai import Crew, Process
+from crewai.memory import LongTermMemory, ShortTermMemory, EntityMemory
+from crewai.memory.storage import LTMSQLiteStorage, RAGStorage
+from typing import List, Optional
 
 # Assemble your crew with memory capabilities
-my_crew = Crew(
-    agents=[...],
-    tasks=[...],
-    process="Process.sequential",
-    memory=True,
-    long_term_memory=EnhanceLongTermMemory(
+my_crew: Crew = Crew(
+    agents = [...],
+    tasks = [...],
+    process = Process.sequential,
+    memory = True,
+    # Long-term memory for persistent storage across sessions
+    long_term_memory = LongTermMemory(
         storage=LTMSQLiteStorage(
-            db_path="/my_data_dir/my_crew1/long_term_memory_storage.db"
+            db_path="/my_crew1/long_term_memory_storage.db"
         )
     ),
-    short_term_memory=EnhanceShortTermMemory(
-        storage=CustomRAGStorage(
-            crew_name="my_crew",
-            storage_type="short_term",
-            data_dir="//my_data_dir",
-            model=embedder["model"],
-            dimension=embedder["dimension"],
+    # Short-term memory for current context using RAG
+    short_term_memory = ShortTermMemory(
+        storage = RAGStorage(
+                embedder_config={
+                    "provider": "openai",
+                    "config": {
+                        "model": 'text-embedding-3-small'
+                    }
+                },
+                type="short_term",
+                path="/my_crew1/"
+            )
         ),
     ),
-    entity_memory=EnhanceEntityMemory(
-        storage=CustomRAGStorage(
-            crew_name="my_crew",
-            storage_type="entities",
-            data_dir="//my_data_dir",
-            model=embedder["model"],
-            dimension=embedder["dimension"],
-        ),
+    # Entity memory for tracking key information about entities
+    entity_memory = EntityMemory(
+        storage=RAGStorage(
+            embedder_config={
+                "provider": "openai",
+                "config": {
+                    "model": 'text-embedding-3-small'
+                }
+            },
+            type="short_term",
+            path="/my_crew1/"
+        )
     ),
     verbose=True,
 )
 ```
 
+## Security Considerations
+
+When configuring memory storage:
+- Use environment variables for storage paths (e.g., `CREWAI_STORAGE_DIR`)
+- Never hardcode sensitive information like database credentials
+- Consider access permissions for storage directories
+- Use relative paths when possible to maintain portability
+
+Example using environment variables:
+```python
+import os
+from crewai import Crew
+from crewai.memory import LongTermMemory
+from crewai.memory.storage import LTMSQLiteStorage
+
+# Configure storage path using environment variable
+storage_path = os.getenv("CREWAI_STORAGE_DIR", "./storage")
+crew = Crew(
+    memory=True,
+    long_term_memory=LongTermMemory(
+        storage=LTMSQLiteStorage(
+            db_path="{storage_path}/memory.db".format(storage_path=storage_path)
+        )
+    )
+)
+```
+
+## Configuration Examples
+
+### Basic Memory Configuration
+```python
+from crewai import Crew
+from crewai.memory import LongTermMemory
+
+# Simple memory configuration
+crew = Crew(memory=True)  # Uses default storage locations
+```
+
+### Custom Storage Configuration
+```python
+from crewai import Crew
+from crewai.memory import LongTermMemory
+from crewai.memory.storage import LTMSQLiteStorage
+
+# Configure custom storage paths
+crew = Crew(
+    memory=True,
+    long_term_memory=LongTermMemory(
+        storage=LTMSQLiteStorage(db_path="./memory.db")
+    )
+)
+```
+
 ## Integrating Mem0 for Enhanced User Memory
 
 [Mem0](https://mem0.ai/) is a self-improving memory layer for LLM applications, enabling personalized AI experiences. 

docs/concepts/tasks.mdx

@@ -268,7 +268,7 @@ analysis_task = Task(
 
 Task guardrails provide a way to validate and transform task outputs before they
 are passed to the next task. This feature helps ensure data quality and provides
-efeedback to agents when their output doesn't meet specific criteria.
+feedback to agents when their output doesn't meet specific criteria.
 
 ### Using Task Guardrails
 

docs/tools/filewritetool.mdx

@@ -8,9 +8,9 @@ icon: file-pen
 
 ## Description
 
-The `FileWriterTool` is a component of the crewai_tools package, designed to simplify the process of writing content to files. 
+The `FileWriterTool` is a component of the crewai_tools package, designed to simplify the process of writing content to files with cross-platform compatibility (Windows, Linux, macOS). 
 It is particularly useful in scenarios such as generating reports, saving logs, creating configuration files, and more. 
-This tool supports creating new directories if they don't exist, making it easier to organize your output.
+This tool handles path differences across operating systems, supports UTF-8 encoding, and automatically creates directories if they don't exist, making it easier to organize your output reliably across different platforms.
 
 ## Installation
 
@@ -43,6 +43,8 @@ print(result)
 
 ## Conclusion
 
-By integrating the `FileWriterTool` into your crews, the agents can execute the process of writing content to files and creating directories. 
-This tool is essential for tasks that require saving output data, creating structured file systems, and more. By adhering to the setup and usage guidelines provided, 
-incorporating this tool into projects is straightforward and efficient.
+By integrating the `FileWriterTool` into your crews, the agents can reliably write content to files across different operating systems. 
+This tool is essential for tasks that require saving output data, creating structured file systems, and handling cross-platform file operations. 
+It's particularly recommended for Windows users who may encounter file writing issues with standard Python file operations.
+
+By adhering to the setup and usage guidelines provided, incorporating this tool into projects is straightforward and ensures consistent file writing behavior across all platforms.

pyproject.toml

@@ -13,6 +13,7 @@ dependencies = [
     "openai>=1.13.3",
     "litellm==1.60.2",
     "instructor>=1.3.3",
+    "timeout-decorator>=0.5.0",
     # Text Processing
     "pdfplumber>=0.11.4",
     "regex>=2024.9.11",

src/crewai/agent.py

@@ -1,9 +1,11 @@
 import re
 import shutil
 import subprocess
+import threading
 from typing import Any, Dict, List, Literal, Optional, Sequence, Union
 
 from pydantic import Field, InstanceOf, PrivateAttr, model_validator
+import timeout_decorator
 
 from crewai.agents import CacheHandler
 from crewai.agents.agent_builder.base_agent import BaseAgent
@@ -16,7 +18,6 @@ from crewai.memory.contextual.contextual_memory import ContextualMemory
 from crewai.task import Task
 from crewai.tools import BaseTool
 from crewai.tools.agent_tools.agent_tools import AgentTools
-from crewai.tools.base_tool import Tool
 from crewai.utilities import Converter, Prompts
 from crewai.utilities.constants import TRAINED_AGENTS_DATA_FILE, TRAINING_DATA_FILE
 from crewai.utilities.converter import generate_model_description
@@ -65,6 +66,7 @@ class Agent(BaseAgent):
     """
 
     _times_executed: int = PrivateAttr(default=0)
+    _have_forced_answer: bool = PrivateAttr(default=False)
     max_execution_time: Optional[int] = Field(
         default=None,
         description="Maximum execution time for an agent to execute a task",
@@ -146,7 +148,7 @@ class Agent(BaseAgent):
     def _set_knowledge(self):
         try:
             if self.knowledge_sources:
-                full_pattern = re.compile(r'[^a-zA-Z0-9\-_\r\n]|(\.\.)')
+                full_pattern = re.compile(r"[^a-zA-Z0-9\-_\r\n]|(\.\.)")
                 knowledge_agent_name = f"{re.sub(full_pattern, '_', self.role)}"
                 if isinstance(self.knowledge_sources, list) and all(
                     isinstance(k, BaseKnowledgeSource) for k in self.knowledge_sources
@@ -160,6 +162,77 @@ class Agent(BaseAgent):
         except (TypeError, ValueError) as e:
             raise ValueError(f"Invalid Knowledge Configuration: {str(e)}")
 
+    def _execute_with_timeout(
+        self,
+        task: Task,
+        context: Optional[str],
+        tools: Optional[List[BaseTool]],
+        timeout: int
+    ) -> str:
+        """Execute task with timeout using thread-based timeout.
+        
+        Args:
+            task: The task to execute
+            context: Optional context for the task
+            tools: Optional list of tools to use
+            timeout: Maximum execution time in seconds (must be > 0)
+            
+        Returns:
+            The result of the task execution, with force_final_answer prompt appended on timeout
+            
+        Raises:
+            ValueError: If timeout is not a positive integer
+            Exception: Any error that occurs during execution
+        """
+        # Validate timeout before creating any resources
+        if not isinstance(timeout, int) or timeout <= 0:
+            raise ValueError("Timeout must be a positive integer greater than zero")
+
+        completion_event: threading.Event = threading.Event()
+        result_container: List[Optional[str]] = [None]
+        error_container: List[Optional[Exception]] = [None]
+        
+        def target() -> None:
+            try:
+                result_container[0] = self._execute_task_without_timeout(task, context, tools)
+            except Exception as e:
+                error_container[0] = e
+            finally:
+                completion_event.set()
+        
+        thread: threading.Thread = threading.Thread(target=target)
+        thread.daemon = True  # Ensures thread doesn't prevent program exit
+        thread.start()
+        
+        # Wait for either completion or timeout
+        completed: bool = completion_event.wait(timeout=timeout)
+        
+        if not completed:
+            self._logger.log("warning", f"Task execution timed out after {timeout} seconds")
+            thread.join(timeout=0.1) 
+            
+            # Clean up resources
+            if hasattr(self, 'agent_executor') and self.agent_executor:
+                self.agent_executor.llm = None  # Release LLM resources
+                if hasattr(self.agent_executor, 'close'):
+                    self.agent_executor.close()
+
+            # Force final answer using the prompt
+            self._have_forced_answer = True
+            forced_answer = self.i18n.errors("force_final_answer")
+            return f"{result_container[0] if result_container[0] else ''}\n{forced_answer}"
+        
+        if error_container[0]:
+            error = error_container[0]
+            self._logger.log("error", f"Task execution failed: {str(error)}")
+            raise error
+        
+        if result_container[0] is None:
+            self._logger.log("warning", "Task execution completed but returned no result")
+            raise timeout_decorator.TimeoutError("Task execution completed but returned no result")  # This is a different kind of failure than timeout
+        
+        return result_container[0]
+
     def execute_task(
         self,
         task: Task,
@@ -175,7 +248,42 @@ class Agent(BaseAgent):
 
         Returns:
             Output of the agent
+
+        Raises:
+            TimeoutError: If the task execution exceeds max_execution_time (if set)
+            Exception: For other execution errors
         """
+        if self.max_execution_time is None:
+            return self._execute_task_without_timeout(task, context, tools)
+
+        original_llm_timeout = getattr(self.llm, 'timeout', None)
+        try:
+            if hasattr(self.llm, 'timeout'):
+                self.llm.timeout = self.max_execution_time
+            
+            result = self._execute_with_timeout(task, context, tools, self.max_execution_time)
+            if self._have_forced_answer:
+                self._logger.log("warning", f"Task '{task.description}' execution timed out after {self.max_execution_time} seconds. Using forced answer.")
+            return result
+        except timeout_decorator.TimeoutError:
+            # This is a different kind of failure (e.g., no result at all)
+            error_msg = (
+                f"Task '{task.description}' execution timed out after {self.max_execution_time} seconds "
+                f"and produced no result. Consider increasing max_execution_time or optimizing the task."
+            )
+            self._logger.log("error", error_msg)
+            raise TimeoutError(error_msg)
+        finally:
+            if original_llm_timeout is not None and hasattr(self.llm, 'timeout'):
+                self.llm.timeout = original_llm_timeout
+
+    def _execute_task_without_timeout(
+        self,
+        task: Task,
+        context: Optional[str] = None,
+        tools: Optional[List[BaseTool]] = None,
+    ) -> str:
+        """Execute task without timeout - contains the original execute_task logic."""
         if self.tools_handler:
             self.tools_handler.last_used_tool = {}  # type: ignore # Incompatible types in assignment (expression has type "dict[Never, Never]", variable has type "ToolCalling")
 

src/crewai/cli/reset_memories_command.py

@@ -3,11 +3,6 @@ import subprocess
 import click
 
 from crewai.cli.utils import get_crew
-from crewai.knowledge.storage.knowledge_storage import KnowledgeStorage
-from crewai.memory.entity.entity_memory import EntityMemory
-from crewai.memory.long_term.long_term_memory import LongTermMemory
-from crewai.memory.short_term.short_term_memory import ShortTermMemory
-from crewai.utilities.task_output_storage_handler import TaskOutputStorageHandler
 
 
 def reset_memories_command(

src/crewai/crew.py

@@ -1,7 +1,6 @@
 import asyncio
 import json
 import re
-import sys
 import uuid
 import warnings
 from concurrent.futures import Future
@@ -381,6 +380,22 @@ class Crew(BaseModel):
 
         return self
 
+    @model_validator(mode="after")
+    def validate_must_have_non_conditional_task(self) -> "Crew":
+        """Ensure that a crew has at least one non-conditional task."""
+        if not self.tasks:
+            return self
+        non_conditional_count = sum(
+            1 for task in self.tasks if not isinstance(task, ConditionalTask)
+        )
+        if non_conditional_count == 0:
+            raise PydanticCustomError(
+                "only_conditional_tasks",
+                "Crew must include at least one non-conditional task",
+                {},
+            )
+        return self
+
     @model_validator(mode="after")
     def validate_first_task(self) -> "Crew":
         """Ensure the first task is not a ConditionalTask."""
@@ -441,6 +456,7 @@ class Crew(BaseModel):
         return self
 
 
+
     @property
     def key(self) -> str:
         source = [agent.key for agent in self.agents] + [
@@ -743,6 +759,7 @@ class Crew(BaseModel):
                     task, task_outputs, futures, task_index, was_replayed
                 )
                 if skipped_task_output:
+                    task_outputs.append(skipped_task_output)
                     continue
 
             if task.async_execution:
@@ -766,7 +783,7 @@ class Crew(BaseModel):
                     context=context,
                     tools=tools_for_task,
                 )
-                task_outputs = [task_output]
+                task_outputs.append(task_output)
                 self._process_task_result(task, task_output)
                 self._store_execution_log(task, task_output, task_index, was_replayed)
 
@@ -787,7 +804,7 @@ class Crew(BaseModel):
             task_outputs = self._process_async_tasks(futures, was_replayed)
             futures.clear()
 
-        previous_output = task_outputs[task_index - 1] if task_outputs else None
+        previous_output = task_outputs[-1] if task_outputs else None
         if previous_output is not None and not task.should_execute(previous_output):
             self._logger.log(
                 "debug",
@@ -909,11 +926,15 @@ class Crew(BaseModel):
             )
 
     def _create_crew_output(self, task_outputs: List[TaskOutput]) -> CrewOutput:
-        if len(task_outputs) != 1:
-            raise ValueError(
-                "Something went wrong. Kickoff should return only one task output."
-            )
-        final_task_output = task_outputs[0]
+        if not task_outputs:
+            raise ValueError("No task outputs available to create crew output.")
+            
+        # Filter out empty outputs and get the last valid one as the main output
+        valid_outputs = [t for t in task_outputs if t.raw]
+        if not valid_outputs:
+            raise ValueError("No valid task outputs available to create crew output.")
+        final_task_output = valid_outputs[-1]
+            
         final_string_output = final_task_output.raw
         self._finish_execution(final_string_output)
         token_usage = self.calculate_usage_metrics()
@@ -922,7 +943,7 @@ class Crew(BaseModel):
             raw=final_task_output.raw,
             pydantic=final_task_output.pydantic,
             json_dict=final_task_output.json_dict,
-            tasks_output=[task.output for task in self.tasks if task.output],
+            tasks_output=task_outputs,
             token_usage=token_usage,
         )
 

src/crewai/llm.py

@@ -164,6 +164,7 @@ class LLM:
         self.context_window_size = 0
         self.reasoning_effort = reasoning_effort
         self.additional_params = kwargs
+        self.is_anthropic = self._is_anthropic_model(model)
 
         litellm.drop_params = True
 
@@ -178,42 +179,62 @@ class LLM:
         self.set_callbacks(callbacks)
         self.set_env_callbacks()
 
+    def _is_anthropic_model(self, model: str) -> bool:
+        """Determine if the model is from Anthropic provider.
+        
+        Args:
+            model: The model identifier string.
+            
+        Returns:
+            bool: True if the model is from Anthropic, False otherwise.
+        """
+        ANTHROPIC_PREFIXES = ('anthropic/', 'claude-', 'claude/')
+        return any(prefix in model.lower() for prefix in ANTHROPIC_PREFIXES)
+
     def call(
         self,
         messages: Union[str, List[Dict[str, str]]],
         tools: Optional[List[dict]] = None,
         callbacks: Optional[List[Any]] = None,
         available_functions: Optional[Dict[str, Any]] = None,
-    ) -> str:
-        """
-        High-level llm call method that:
-          1) Accepts either a string or a list of messages
-          2) Converts string input to the required message format
-          3) Calls litellm.completion
-          4) Handles function/tool calls if any
-          5) Returns the final text response or tool result
-
-        Parameters:
-        - messages (Union[str, List[Dict[str, str]]]): The input messages for the LLM.
-          - If a string is provided, it will be converted into a message list with a single entry.
-          - If a list of dictionaries is provided, each dictionary should have 'role' and 'content' keys.
-        - tools (Optional[List[dict]]): A list of tool schemas for function calling.
-        - callbacks (Optional[List[Any]]): A list of callback functions to be executed.
-        - available_functions (Optional[Dict[str, Any]]): A dictionary mapping function names to actual Python functions.
-
+    ) -> Union[str, Any]:
+        """High-level LLM call method.
+        
+        Args:
+            messages: Input messages for the LLM.
+                     Can be a string or list of message dictionaries.
+                     If string, it will be converted to a single user message.
+                     If list, each dict must have 'role' and 'content' keys.
+            tools: Optional list of tool schemas for function calling.
+                  Each tool should define its name, description, and parameters.
+            callbacks: Optional list of callback functions to be executed
+                      during and after the LLM call.
+            available_functions: Optional dict mapping function names to callables
+                               that can be invoked by the LLM.
+        
         Returns:
-        - str: The final text response from the LLM or the result of a tool function call.
-
+            Union[str, Any]: Either a text response from the LLM (str) or
+                           the result of a tool function call (Any).
+        
+        Raises:
+            TypeError: If messages format is invalid
+            ValueError: If response format is not supported
+            LLMContextLengthExceededException: If input exceeds model's context limit
+        
         Examples:
-        ---------
-        # Example 1: Using a string input
-        response = llm.call("Return the name of a random city in the world.")
-        print(response)
-
-        # Example 2: Using a list of messages
-        messages = [{"role": "user", "content": "What is the capital of France?"}]
-        response = llm.call(messages)
-        print(response)
+            # Example 1: Simple string input
+            >>> response = llm.call("Return the name of a random city.")
+            >>> print(response)
+            "Paris"
+            
+            # Example 2: Message list with system and user messages
+            >>> messages = [
+            ...     {"role": "system", "content": "You are a geography expert"},
+            ...     {"role": "user", "content": "What is France's capital?"}
+            ... ]
+            >>> response = llm.call(messages)
+            >>> print(response)
+            "The capital of France is Paris."
         """
         # Validate parameters before proceeding with the call.
         self._validate_call_params()
@@ -233,10 +254,13 @@ class LLM:
                 self.set_callbacks(callbacks)
 
             try:
-                # --- 1) Prepare the parameters for the completion call
+                # --- 1) Format messages according to provider requirements
+                formatted_messages = self._format_messages_for_provider(messages)
+
+                # --- 2) Prepare the parameters for the completion call
                 params = {
                     "model": self.model,
-                    "messages": messages,
+                    "messages": formatted_messages,
                     "timeout": self.timeout,
                     "temperature": self.temperature,
                     "top_p": self.top_p,
@@ -324,6 +348,38 @@ class LLM:
                     logging.error(f"LiteLLM call failed: {str(e)}")
                 raise
 
+    def _format_messages_for_provider(self, messages: List[Dict[str, str]]) -> List[Dict[str, str]]:
+        """Format messages according to provider requirements.
+        
+        Args:
+            messages: List of message dictionaries with 'role' and 'content' keys.
+                     Can be empty or None.
+        
+        Returns:
+            List of formatted messages according to provider requirements.
+            For Anthropic models, ensures first message has 'user' role.
+        
+        Raises:
+            TypeError: If messages is None or contains invalid message format.
+        """
+        if messages is None:
+            raise TypeError("Messages cannot be None")
+            
+        # Validate message format first
+        for msg in messages:
+            if not isinstance(msg, dict) or "role" not in msg or "content" not in msg:
+                raise TypeError("Invalid message format. Each message must be a dict with 'role' and 'content' keys")
+            
+        if not self.is_anthropic:
+            return messages
+            
+        # Anthropic requires messages to start with 'user' role
+        if not messages or messages[0]["role"] == "system":
+            # If first message is system or empty, add a placeholder user message
+            return [{"role": "user", "content": "."}, *messages]
+                
+        return messages
+
     def _get_custom_llm_provider(self) -> str:
         """
         Derives the custom_llm_provider from the model string.

src/crewai/memory/entity/entity_memory.py

@@ -1,4 +1,4 @@
-from typing import Any, Optional
+from typing import Optional
 
 from pydantic import PrivateAttr
 

src/crewai/memory/memory.py

@@ -1,9 +1,7 @@
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional
 
 from pydantic import BaseModel
 
-from crewai.memory.storage.rag_storage import RAGStorage
-
 
 class Memory(BaseModel):
     """

src/crewai/task.py

@@ -674,19 +674,32 @@ class Task(BaseModel):
             return OutputFormat.PYDANTIC
         return OutputFormat.RAW
 
-    def _save_file(self, result: Any) -> None:
+    def _save_file(self, result: Union[Dict, str, Any]) -> None:
         """Save task output to a file.
 
+        Note:
+            For cross-platform file writing, especially on Windows, consider using FileWriterTool
+            from the crewai_tools package:
+                pip install 'crewai[tools]'
+                from crewai_tools import FileWriterTool
+
         Args:
             result: The result to save to the file. Can be a dict or any stringifiable object.
 
         Raises:
             ValueError: If output_file is not set
-            RuntimeError: If there is an error writing to the file
+            RuntimeError: If there is an error writing to the file. For cross-platform
+                compatibility, especially on Windows, use FileWriterTool from crewai_tools
+                package.
         """
         if self.output_file is None:
             raise ValueError("output_file is not set.")
 
+        FILEWRITER_RECOMMENDATION = (
+            "For cross-platform file writing, especially on Windows, "
+            "use FileWriterTool from crewai_tools package."
+        )
+
         try:
             resolved_path = Path(self.output_file).expanduser().resolve()
             directory = resolved_path.parent
@@ -702,7 +715,12 @@ class Task(BaseModel):
                 else:
                     file.write(str(result))
         except (OSError, IOError) as e:
-            raise RuntimeError(f"Failed to save output file: {e}")
+            raise RuntimeError(
+                "\n".join([
+                    f"Failed to save output file: {e}",
+                    FILEWRITER_RECOMMENDATION
+                ])
+            )
         return None
 
     def __repr__(self):

tests/crew_test.py

@@ -49,6 +49,39 @@ writer = Agent(
 )
 
 
+def test_crew_with_only_conditional_tasks_raises_error():
+    """Test that creating a crew with only conditional tasks raises an error."""
+    def condition_func(task_output: TaskOutput) -> bool:
+        return True
+
+    conditional1 = ConditionalTask(
+        description="Conditional task 1",
+        expected_output="Output 1",
+        agent=researcher,
+        condition=condition_func,
+    )
+    conditional2 = ConditionalTask(
+        description="Conditional task 2",
+        expected_output="Output 2",
+        agent=researcher,
+        condition=condition_func,
+    )
+    conditional3 = ConditionalTask(
+        description="Conditional task 3",
+        expected_output="Output 3",
+        agent=researcher,
+        condition=condition_func,
+    )
+
+    with pytest.raises(
+        pydantic_core._pydantic_core.ValidationError,
+        match="Crew must include at least one non-conditional task",
+    ):
+        Crew(
+            agents=[researcher],
+            tasks=[conditional1, conditional2, conditional3],
+        )
+
 def test_crew_config_conditional_requirement():
     with pytest.raises(ValueError):
         Crew(process=Process.sequential)
@@ -2060,6 +2093,195 @@ def test_tools_with_custom_caching():
             assert result.raw == "3"
 
 
+@pytest.mark.vcr(filter_headers=["authorization"])
+def test_conditional_task_uses_last_output():
+    """Test that conditional tasks use the last task output for condition evaluation."""
+    task1 = Task(
+        description="First task",
+        expected_output="First output",
+        agent=researcher,
+    )
+    def condition_fails(task_output: TaskOutput) -> bool:
+        # This condition will never be met
+        return "never matches" in task_output.raw.lower()
+    
+    def condition_succeeds(task_output: TaskOutput) -> bool:
+        # This condition will match first task's output
+        return "first success" in task_output.raw.lower()
+    
+    conditional_task1 = ConditionalTask(
+        description="Second task - conditional that fails condition",
+        expected_output="Second output",
+        agent=researcher,
+        condition=condition_fails,
+    )
+    
+    conditional_task2 = ConditionalTask(
+        description="Third task - conditional that succeeds using first task output",
+        expected_output="Third output",
+        agent=writer,
+        condition=condition_succeeds,
+    )
+
+    crew = Crew(
+        agents=[researcher, writer],
+        tasks=[task1, conditional_task1, conditional_task2],
+    )
+
+    # Mock outputs for tasks
+    mock_first = TaskOutput(
+        description="First task output",
+        raw="First success output",  # Will be used by third task's condition
+        agent=researcher.role,
+    )
+    mock_skipped = TaskOutput(
+        description="Second task output",
+        raw="",  # Empty output since condition fails
+        agent=researcher.role,
+    )
+    mock_third = TaskOutput(
+        description="Third task output",
+        raw="Third task executed",  # Output when condition succeeds using first task output
+        agent=writer.role,
+    )
+    
+    # Set up mocks for task execution and conditional logic
+    with patch.object(ConditionalTask, "should_execute") as mock_should_execute:
+        # First conditional fails, second succeeds
+        mock_should_execute.side_effect = [False, True]
+        
+        with patch.object(Task, "execute_sync") as mock_execute:
+            mock_execute.side_effect = [mock_first, mock_third]
+            result = crew.kickoff()
+            
+            # Verify execution behavior
+            assert mock_execute.call_count == 2  # Only first and third tasks execute
+            assert mock_should_execute.call_count == 2  # Both conditionals checked
+            
+            # Verify outputs collection
+            assert len(result.tasks_output) == 3
+            assert result.tasks_output[0].raw == "First success output"  # First task succeeded
+            assert result.tasks_output[1].raw == ""  # Second task skipped (condition failed)
+            assert result.tasks_output[2].raw == "Third task executed"  # Third task used first task's output
+
+@pytest.mark.vcr(filter_headers=["authorization"])
+def test_conditional_tasks_result_collection():
+    """Test that task outputs are properly collected based on execution status."""
+    task1 = Task(
+        description="Normal task that always executes",
+        expected_output="First output",
+        agent=researcher,
+    )
+    
+    def condition_never_met(task_output: TaskOutput) -> bool:
+        return "never matches" in task_output.raw.lower()
+    
+    def condition_always_met(task_output: TaskOutput) -> bool:
+        return "success" in task_output.raw.lower()
+    
+    task2 = ConditionalTask(
+        description="Conditional task that never executes",
+        expected_output="Second output",
+        agent=researcher,
+        condition=condition_never_met,
+    )
+    
+    task3 = ConditionalTask(
+        description="Conditional task that always executes",
+        expected_output="Third output",
+        agent=writer,
+        condition=condition_always_met,
+    )
+
+    crew = Crew(
+        agents=[researcher, writer],
+        tasks=[task1, task2, task3],
+    )
+
+    # Mock outputs for different execution paths
+    mock_success = TaskOutput(
+        description="Success output",
+        raw="Success output",  # Triggers third task's condition
+        agent=researcher.role,
+    )
+    mock_skipped = TaskOutput(
+        description="Skipped output",
+        raw="",  # Empty output for skipped task
+        agent=researcher.role,
+    )
+    mock_conditional = TaskOutput(
+        description="Conditional output",
+        raw="Conditional task executed",
+        agent=writer.role,
+    )
+    
+    # Set up mocks for task execution and conditional logic
+    with patch.object(ConditionalTask, "should_execute") as mock_should_execute:
+        # First conditional fails, second succeeds
+        mock_should_execute.side_effect = [False, True]
+        
+        with patch.object(Task, "execute_sync") as mock_execute:
+            mock_execute.side_effect = [mock_success, mock_conditional]
+            result = crew.kickoff()
+            
+            # Verify execution behavior
+            assert mock_execute.call_count == 2  # Only first and third tasks execute
+            assert mock_should_execute.call_count == 2  # Both conditionals checked
+            
+            # Verify task output collection
+            assert len(result.tasks_output) == 3
+            assert result.tasks_output[0].raw == "Success output"      # Normal task executed
+            assert result.tasks_output[1].raw == ""                    # Second task skipped
+            assert result.tasks_output[2].raw == "Conditional task executed"  # Third task executed
+
+@pytest.mark.vcr(filter_headers=["authorization"])
+def test_multiple_conditional_tasks():
+    """Test that having multiple conditional tasks in sequence works correctly."""
+    task1 = Task(
+        description="Initial research task",
+        expected_output="Research output",
+        agent=researcher,
+    )
+    
+    def condition1(task_output: TaskOutput) -> bool:
+        return "success" in task_output.raw.lower()
+    
+    def condition2(task_output: TaskOutput) -> bool:
+        return "proceed" in task_output.raw.lower()
+    
+    task2 = ConditionalTask(
+        description="First conditional task",
+        expected_output="Conditional output 1",
+        agent=writer,
+        condition=condition1,
+    )
+    
+    task3 = ConditionalTask(
+        description="Second conditional task",
+        expected_output="Conditional output 2",
+        agent=writer,
+        condition=condition2,
+    )
+
+    crew = Crew(
+        agents=[researcher, writer],
+        tasks=[task1, task2, task3],
+    )
+
+    # Mock different task outputs to test conditional logic
+    mock_success = TaskOutput(
+        description="Mock success",
+        raw="Success and proceed output",
+        agent=researcher.role,
+    )
+    
+    # Set up mocks for task execution
+    with patch.object(Task, "execute_sync", return_value=mock_success) as mock_execute:
+        result = crew.kickoff()
+        # Verify all tasks were executed (no IndexError)
+        assert mock_execute.call_count == 3
+        assert len(result.tasks_output) == 3
+
 @pytest.mark.vcr(filter_headers=["authorization"])
 def test_using_contextual_memory():
     from unittest.mock import patch

tests/llm_test.py

@@ -286,6 +286,79 @@ def test_o3_mini_reasoning_effort_medium():
 
 
 @pytest.mark.vcr(filter_headers=["authorization"])
+@pytest.fixture
+def anthropic_llm():
+    """Fixture providing an Anthropic LLM instance."""
+    return LLM(model="anthropic/claude-3-sonnet")
+
+@pytest.fixture
+def system_message():
+    """Fixture providing a system message."""
+    return {"role": "system", "content": "test"}
+
+@pytest.fixture
+def user_message():
+    """Fixture providing a user message."""
+    return {"role": "user", "content": "test"}
+
+def test_anthropic_message_formatting_edge_cases(anthropic_llm):
+    """Test edge cases for Anthropic message formatting."""
+    # Test None messages
+    with pytest.raises(TypeError, match="Messages cannot be None"):
+        anthropic_llm._format_messages_for_provider(None)
+        
+    # Test empty message list
+    formatted = anthropic_llm._format_messages_for_provider([])
+    assert len(formatted) == 1
+    assert formatted[0]["role"] == "user"
+    assert formatted[0]["content"] == "."
+    
+    # Test invalid message format
+    with pytest.raises(TypeError, match="Invalid message format"):
+        anthropic_llm._format_messages_for_provider([{"invalid": "message"}])
+
+def test_anthropic_model_detection():
+    """Test Anthropic model detection with various formats."""
+    models = [
+        ("anthropic/claude-3", True),
+        ("claude-instant", True),
+        ("claude/v1", True),
+        ("gpt-4", False),
+        ("", False),
+        ("anthropomorphic", False),  # Should not match partial words
+    ]
+    
+    for model, expected in models:
+        llm = LLM(model=model)
+        assert llm.is_anthropic == expected, f"Failed for model: {model}"
+
+def test_anthropic_message_formatting(anthropic_llm, system_message, user_message):
+    """Test Anthropic message formatting with fixtures."""
+    # Test when first message is system
+    formatted = anthropic_llm._format_messages_for_provider([system_message])
+    assert len(formatted) == 2
+    assert formatted[0]["role"] == "user"
+    assert formatted[0]["content"] == "."
+    assert formatted[1] == system_message
+
+    # Test when first message is already user
+    formatted = anthropic_llm._format_messages_for_provider([user_message])
+    assert len(formatted) == 1
+    assert formatted[0] == user_message
+
+    # Test with empty message list
+    formatted = anthropic_llm._format_messages_for_provider([])
+    assert len(formatted) == 1
+    assert formatted[0]["role"] == "user"
+    assert formatted[0]["content"] == "."
+
+    # Test with non-Anthropic model (should not modify messages)
+    non_anthropic_llm = LLM(model="gpt-4")
+    formatted = non_anthropic_llm._format_messages_for_provider([system_message])
+    assert len(formatted) == 1
+    assert formatted[0] == system_message
+
+
 def test_deepseek_r1_with_open_router():
     if not os.getenv("OPEN_ROUTER_API_KEY"):
         pytest.skip("OPEN_ROUTER_API_KEY not set; skipping test.")

tests/task_test.py

@@ -13,6 +13,7 @@ from crewai import Agent, Crew, Process, Task
 from crewai.tasks.conditional_task import ConditionalTask
 from crewai.tasks.task_output import TaskOutput
 from crewai.utilities.converter import Converter
+from concurrent.futures import TimeoutError as FuturesTimeoutError
 
 
 def test_task_tool_reflect_agent_tools():
@@ -1283,3 +1284,139 @@ def test_interpolate_valid_types():
     assert parsed["optional"] is None
     assert parsed["nested"]["flag"] is True
     assert parsed["nested"]["empty"] is None
+
+
+def test_task_completes_within_max_execution_time():
+    """Test task completes successfully within specified timeout"""
+    researcher = Agent(
+        role="Researcher",
+        goal="Test goal",
+        backstory="Test backstory",
+        max_execution_time=100  # Ample time for completion
+    )
+
+    task = Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=researcher
+    )
+
+    with patch.object(Agent, '_execute_task_without_timeout', return_value="Success") as mock_execute:
+        result = task.execute_sync(agent=researcher)
+        assert result.raw == "Success"
+        mock_execute.assert_called_once()
+
+
+def test_task_exceeds_max_execution_time():
+    """Test task raises TimeoutError when exceeding max execution time"""
+    researcher = Agent(
+        role="Researcher",
+        goal="Test goal",
+        backstory="Test backstory",
+        max_execution_time=1  # Very short timeout
+    )
+
+    task = Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=researcher
+    )
+
+    # Mock the OpenAI API call to avoid authentication
+    with patch('litellm.completion', side_effect=TimeoutError("Request timed out")):
+        with pytest.raises(TimeoutError) as excinfo:
+            task.execute_sync(agent=researcher)
+        assert "timed out" in str(excinfo.value)
+
+
+def test_task_no_max_execution_time():
+    """Test task executes normally without timeout setting"""
+    researcher = Agent(
+        role="Researcher",
+        goal="Test goal",
+        backstory="Test backstory",
+        max_execution_time=None  # No timeout
+    )
+
+    task = Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=researcher
+    )
+
+    with patch.object(Agent, 'execute_task', return_value="Success") as mock_execute:
+        result = task.execute_sync(agent=researcher)
+        assert result.raw == "Success"
+        mock_execute.assert_called_once()
+
+
+def test_task_max_execution_time_zero():
+    """Test immediate timeout with max_execution_time=0"""
+    researcher = Agent(
+        role="Researcher",
+        goal="Test goal",
+        backstory="Test backstory",
+        max_execution_time=1  # Set to minimum valid value
+    )
+
+    task = Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=researcher
+    )
+
+    # Simulate immediate timeout using FuturesTimeoutError
+    with patch('concurrent.futures.ThreadPoolExecutor') as mock_executor:
+        mock_future = MagicMock()
+        mock_future.result.side_effect = FuturesTimeoutError()
+        mock_executor.return_value.submit.return_value = mock_future
+
+        with pytest.raises(TimeoutError) as excinfo:
+            task.execute_sync(agent=researcher)
+        assert "timed out after 1 seconds" in str(excinfo.value)
+
+
+def test_task_force_final_answer_on_timeout():
+    """Test that force_final_answer is used when task times out"""
+    researcher = Agent(
+        role="Researcher",
+        goal="Test goal",
+        backstory="Test backstory",
+        max_execution_time=1  # Very short timeout
+    )
+
+    task = Task(
+        description="Test task",
+        expected_output="Test output",
+        agent=researcher
+    )
+
+    # Mock the task execution to simulate a partial result before timeout
+    mock_i18n = MagicMock()
+    mock_i18n.errors.return_value = "MUST give your absolute best final answer"
+    researcher.i18n = mock_i18n
+
+    class MockThread:
+        def __init__(self, target, *args, **kwargs):
+            self.target = target
+            self.daemon = kwargs.get('daemon', False)
+            self.args = args
+            self.kwargs = kwargs
+
+        def start(self):
+            # Execute the target function to set the result
+            self.target()
+
+        def join(self, timeout=None):
+            pass
+
+    def mock_thread(*args, **kwargs):
+        return MockThread(*args, **kwargs)
+
+    with patch('threading.Thread', side_effect=mock_thread), \
+         patch('threading.Event.wait', return_value=False), \
+         patch('litellm.completion'), \
+         patch.object(Agent, '_execute_task_without_timeout', return_value="Partial result"):
+        result = task.execute_sync(agent=researcher)
+        assert "MUST give your absolute best final answer" in result.raw
+        assert "Partial result" in result.raw  # Should include partial result