* Changed v1.1 -> v2
* Fixed Test Cases:
* Fixed linting issues
* Changed docs
* Refractored the storage
* Fixed test cases
* Fixing run-time checks
* Fixed Test Case
* Updated docs and added test case for custom categories
* Add the TODO back
* Minor Changes
* Added output_format in search
* Minor changes
* Added output_format and version in both search and save
* Small change
* Minor bugs
* Fixed test cases
* Changed docs
---------
Co-authored-by: Lucas Gomide <lucaslg200@gmail.com>
* feat: add exchanged messages in LLMCallCompletedEvent
* feat: add GoalAlignment metric for Agent evaluation
* feat: add SemanticQuality metric for Agent evaluation
* feat: add Tool Metrics for Agent evaluation
* feat: add Reasoning Metrics for Agent evaluation, still in progress
* feat: add AgentEvaluator class
This class will evaluate Agent' results and report to user
* fix: do not evaluate Agent by default
This is a experimental feature we still need refine it further
* test: add Agent eval tests
* fix: render all feedback per iteration
* style: resolve linter issues
* style: fix mypy issues
* fix: allow messages be empty on LLMCallCompletedEvent
* feat: add Experiment evaluation framework with baseline comparison
* fix: reset evaluator for each experiement iteraction
* fix: fix track of new test cases
* chore: split Experimental evaluation classes
* refactor: remove unused method
* refactor: isolate Console print in a dedicated class
* fix: make crew required to run an experiment
* fix: use time-aware to define experiment result
* test: add tests for Evaluator Experiment
* style: fix linter issues
* fix: encode string before hashing
* style: resolve linter issues
* feat: add experimental folder for beta features (#3141)
* test: move tests to experimental folder
* Fix#3149: Add missing create_directory parameter to Task class
- Add create_directory field with default value True for backward compatibility
- Update _save_file method to respect create_directory parameter
- Add comprehensive tests covering all scenarios
- Maintain existing behavior when create_directory=True (default)
The create_directory parameter was documented but missing from implementation.
Users can now control directory creation behavior:
- create_directory=True (default): Creates directories if they don't exist
- create_directory=False: Raises RuntimeError if directory doesn't exist
Fixes issue where users got TypeError when trying to use the documented
create_directory parameter.
Co-Authored-By: Jo\u00E3o <joao@crewai.com>
* Fix lint: Remove unused import os from test_create_directory_true
- Removes F401 lint error: 'os' imported but unused
- All lint checks should now pass
Co-Authored-By: Jo\u00E3o <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Jo\u00E3o <joao@crewai.com>
* Compaing BaseLLM class instead of LLM
* Fixed test cases
* Fixed Linting Issues
* removed last line
---------
Co-authored-by: Lucas Gomide <lucaslg200@gmail.com>
* Added add_sources()
* Fixed the agent knowledge querying
* Added test cases
* Fixed linting issue
* Fixed logic
* Seems like a falky test case
* Minor changes
* Added knowledge attriute to the crew documentation
* Flaky test
* fixed spaces
* Flaky Test Case
* Seems like a flaky test case
---------
Co-authored-by: Lucas Gomide <lucaslg200@gmail.com>
* feat: add exchanged messages in LLMCallCompletedEvent
* feat: add GoalAlignment metric for Agent evaluation
* feat: add SemanticQuality metric for Agent evaluation
* feat: add Tool Metrics for Agent evaluation
* feat: add Reasoning Metrics for Agent evaluation, still in progress
* feat: add AgentEvaluator class
This class will evaluate Agent' results and report to user
* fix: do not evaluate Agent by default
This is a experimental feature we still need refine it further
* test: add Agent eval tests
* fix: render all feedback per iteration
* style: resolve linter issues
* style: fix mypy issues
* fix: allow messages be empty on LLMCallCompletedEvent
Add crew context tracking using OpenTelemetry baggage for thread-safe propagation. Context is set during kickoff and cleaned up in finally block. Added thread safety tests with mocked agent execution.
* fix: clean up whitespace and update dependencies
* Removed unnecessary whitespace in multiple files for consistency.
* Updated `crewai-tools` dependency version to `0.49.0` in `pyproject.toml` and related template files.
* Bumped CrewAI version to `0.140.0` in `__init__.py` for alignment with updated dependencies.
* chore: update pyproject.toml to exclude documentation from build targets
* Added exclusions for the `docs` directory in both wheel and sdist build targets to streamline the build process and reduce unnecessary file inclusion.
* chore: update uv.lock for dependency resolution and Python version compatibility
* Incremented revision to 2.
* Updated resolution markers to include support for Python 3.13 and adjusted platform checks for better compatibility.
* Added new wheel URLs for zstandard version 0.23.0 to ensure availability across various platforms.
* chore: pin json-repair dependency version in pyproject.toml and uv.lock
* Updated json-repair dependency from a range to a specific version (0.25.2) for consistency and to avoid potential compatibility issues.
* Adjusted related entries in uv.lock to reflect the pinned version, ensuring alignment across project files.
* chore: pin agentops dependency version in pyproject.toml and uv.lock
* Updated agentops dependency from a range to a specific version (0.3.18) for consistency and to avoid potential compatibility issues.
* Adjusted related entries in uv.lock to reflect the pinned version, ensuring alignment across project files.
* test: enhance cache call assertions in crew tests
* Improved the test for cache hitting between agents by filtering mock calls to ensure they include the expected 'tool' and 'input' keywords.
* Added assertions to verify the number of cache calls and their expected arguments, enhancing the reliability of the test.
* Cleaned up whitespace and improved readability in various test cases for better maintainability.
* feat: add capability to track LLM calls by task and agent
This makes it possible to filter or scope LLM events by specific agents or tasks, which can be very useful for debugging or analytics in real-time application
* feat: add docs about LLM tracking by Agents and Tasks
* fix incompatible BaseLLM.call method signature
* feat: support to filter LLM Events from Lite Agent
When creating a Crew via the CLI and selecting the Azure provider, the generated .env file had environment variables in lowercase.
This commit ensures that all environment variables are written in uppercase.
* fix: normalize project names by stripping trailing slashes in crew creation
- Strip trailing slashes from project names in create_folder_structure
- Add comprehensive tests for trailing slash scenarios
- Fixes#3059
The issue occurred because trailing slashes in project names like 'hello/'
were directly incorporated into pyproject.toml, creating invalid package
names and script entries. This fix silently normalizes project names by
stripping trailing slashes before processing, maintaining backward
compatibility while fixing the invalid template generation.
Co-Authored-By: João <joao@crewai.com>
* trigger CI re-run to check for flaky test issue
Co-Authored-By: João <joao@crewai.com>
* fix: resolve circular import in CLI authentication module
- Move ToolCommand import to be local inside _poll_for_token method
- Update test mock to patch ToolCommand at correct location
- Resolves Python 3.11 test collection failure in CI
Co-Authored-By: João <joao@crewai.com>
* feat: add comprehensive class name validation for Python identifiers
- Ensure generated class names are always valid Python identifiers
- Handle edge cases: names starting with numbers, special characters, keywords, built-ins
- Add sanitization logic to remove invalid characters and prefix with 'Crew' when needed
- Add comprehensive test coverage for class name validation edge cases
- Addresses GitHub PR comment from lucasgomide about class name validity
Fixes include:
- Names starting with numbers: '123project' -> 'Crew123Project'
- Python built-ins: 'True' -> 'TrueCrew', 'False' -> 'FalseCrew'
- Special characters: 'hello@world' -> 'HelloWorld'
- Empty/whitespace: ' ' -> 'DefaultCrew'
- All generated class names pass isidentifier() and keyword checks
Co-Authored-By: João <joao@crewai.com>
* refactor: change class name validation to raise errors instead of generating defaults
- Remove default value generation (Crew prefix/suffix, DefaultCrew fallback)
- Raise ValueError with descriptive messages for invalid class names
- Update tests to expect validation errors instead of default corrections
- Addresses GitHub comment feedback from lucasgomide about strict validation
Co-Authored-By: João <joao@crewai.com>
* fix: add working directory safety checks to prevent test interference
Co-Authored-By: João <joao@crewai.com>
* fix: standardize working directory handling in tests to prevent corruption
Co-Authored-By: João <joao@crewai.com>
* fix: eliminate os.chdir() usage in tests to prevent working directory corruption
- Replace os.chdir() with parent_folder parameter for create_folder_structure tests
- Mock create_folder_structure directly for create_crew tests to avoid directory changes
- All 12 tests now pass locally without working directory corruption
- Should resolve the 103 failing tests in Python 3.12 CI
Co-Authored-By: João <joao@crewai.com>
* fix: remove unused os import to resolve lint failure
- Remove unused 'import os' statement from test_create_crew.py
- All tests still pass locally after removing unused import
- Should resolve F401 lint error in CI
Co-Authored-By: João <joao@crewai.com>
* feat: add folder name validation for Python module names
- Implement validation to ensure folder_name is valid Python identifier
- Check that folder names don't start with digits
- Validate folder names are not Python keywords
- Sanitize invalid characters from folder names
- Raise ValueError with descriptive messages for invalid cases
- Update tests to validate both folder and class name requirements
- Addresses GitHub comment requiring folder names to be valid Python module names
Co-Authored-By: João <joao@crewai.com>
* fix: correct folder name validation logic to match test expectations
- Fix validation regex to catch names starting with invalid characters like '@#/'
- Ensure validation properly raises ValueError for cases expected by tests
- Maintain support for valid cases like 'my.project/' -> 'myproject'
- Address lucasgomide's comment about valid Python module names
Co-Authored-By: João <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: João <joao@crewai.com>
Co-authored-by: Lucas Gomide <lucaslg200@gmail.com>
* test: add tests to test get_crews
* feat: improve Crew search while resetting their memories
Some memories couldn't be reset due to their reliance on relative external sources like `PDFKnowledge`. This was caused by the need to run the reset memories command from the `src` directory, which could break when external files weren't accessible from that path.
This commit allows the reset command to be executed from the root of the project — the same location typically used to run a crew — improving compatibility and reducing friction.
* feat: skip cli/templates folder while looking for Crew
* refactor: use console.print instead of print
* fix: possible fix for Thinking stuck
* feat: add agent logging events for execution tracking
- Introduced AgentLogsStartedEvent and AgentLogsExecutionEvent to enhance logging capabilities during agent execution.
- Updated CrewAgentExecutor to emit these events at the start and during execution, respectively.
- Modified EventListener to handle the new logging events and format output accordingly in the console.
- Enhanced ConsoleFormatter to display agent logs in a structured format, improving visibility of agent actions and outputs.
* drop emoji
* refactor: improve code structure and logging in LiteAgent and ConsoleFormatter
- Refactored imports in lite_agent.py for better readability.
- Enhanced guardrail property initialization in LiteAgent.
- Updated logging functionality to emit AgentLogsExecutionEvent for better tracking.
- Modified ConsoleFormatter to include tool arguments and final output in status updates.
- Improved output formatting for long text in ConsoleFormatter.
* fix tests
---------
Co-authored-by: Eduardo Chiarotti <dudumelgaco@hotmail.com>
* Fix issue 2993: Prevent Flow status logs from hiding human input
- Add pause_live_updates() and resume_live_updates() methods to ConsoleFormatter
- Modify _ask_human_input() to pause Flow status updates during human input
- Add comprehensive tests for pause/resume functionality and integration
- Ensure Live session is properly managed during human input prompts
- Fix prevents Flow status logs from overwriting user input prompts
Fixes#2993
Co-Authored-By: João <joao@crewai.com>
* Fix lint: Remove unused pytest import
- Remove unused pytest import from test_console_formatter_pause_resume.py
- Fixes F401 lint error identified in CI
Co-Authored-By: João <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: João <joao@crewai.com>
* Fix telemetry singleton pattern to respect dynamic environment variables
- Modified Telemetry.__init__ to prevent re-initialization with _initialized flag
- Updated _safe_telemetry_operation to check _is_telemetry_disabled() dynamically
- Added comprehensive tests for environment variables set after singleton creation
- Fixed singleton contamination in existing tests by adding proper reset
- Resolves issue #2945 where CREWAI_DISABLE_TELEMETRY=true was ignored when set after import
Co-Authored-By: João <joao@crewai.com>
* Implement code review improvements
- Move _initialized flag to __new__ method for better encapsulation
- Add type hints to _safe_telemetry_operation method
- Consolidate telemetry execution checks into _should_execute_telemetry helper
- Add pytest fixtures to reduce test setup redundancy
- Enhanced documentation for singleton behavior
Co-Authored-By: João <joao@crewai.com>
* Fix mypy type-checker errors
- Add explicit bool type annotation to _initialized field
- Fix return value in task_started method to not return _safe_telemetry_operation result
- Simplify initialization logic to set _initialized once in __init__
Co-Authored-By: João <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: João <joao@crewai.com>
Co-authored-by: Lorenze Jay <63378463+lorenzejay@users.noreply.github.com>
* feat: add guardrail support for Agents when using direct kickoff calls
* refactor: expose guardrail func in a proper utils file
* fix: resolve Self import on python 3.10
* test: fix structured tool tests
No tests were being executed from this file
* feat: support to run async tool
Some Tool requires async execution. This commit allow us to collect tool result from coroutines
* docs: add docs about asynchronous tool support
* docs: add organization management in our CLI docs
* feat: improve user feedback when user is not authenticated
* feat: improve logging about current organization while publishing/install a Tool
* feat: improve logging when Agent repository is not found during fetch
* fix linter offences
* test: fix auth token error
* feat: support to list, switch and see your current organization
* feat: store the current org after logged in
* feat: filtering agents, tools and their actions by organization_uuid if present
* fix linter offenses
* refactor: propagate the current org thought Header instead of params
* refactor: rename org column name to ID instead of Handle
---------
Co-authored-by: Tony Kipkemboi <iamtonykipkemboi@gmail.com>
Previously, we only supported tools from the crewai-tools open-source repository. Now, we're introducing improved support for private tool repositories.
* feat: add capability to see and expose public Tool classes
* feat: persist available Tools from repository on publish
* ci: ignore explictly templates from ruff check
Ruff only applies --exclude to files it discovers itself. So we have to skip manually the same files excluded from `ruff.toml`
* sytle: fix linter issues
* refactor: renaming available_tools_classes by available_exports
* feat: provide more context about exportable tools
* feat: allow to install a Tool from pypi
* test: fix tests
* feat: add env_vars attribute to BaseTool
* remove TODO: security check since we are handle that on enterprise side
* Add usage limit feature to BaseTool class
- Add max_usage_count and current_usage_count attributes to BaseTool
- Implement usage limit checking in ToolUsage._use method
- Add comprehensive tests for usage limit functionality
- Maintain backward compatibility with None default for unlimited usage
Co-Authored-By: Joe Moura <joao@crewai.com>
* Fix CI failures and address code review feedback
- Add max_usage_count/current_usage_count to CrewStructuredTool
- Add input validation for positive max_usage_count
- Add reset_usage_count method to BaseTool
- Extract usage limit check into separate method
- Add comprehensive edge case tests
- Add proper type hints throughout
- Fix linting issues
Co-Authored-By: Joe Moura <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Joe Moura <joao@crewai.com>
* feat: Add inject_date flag to Agent for automatic date injection
Co-Authored-By: Joe Moura <joao@crewai.com>
* feat: Add date_format parameter and error handling to inject_date feature
Co-Authored-By: Joe Moura <joao@crewai.com>
* fix: Update test implementation for inject_date feature
Co-Authored-By: Joe Moura <joao@crewai.com>
* fix: Add date format validation to prevent invalid formats
Co-Authored-By: Joe Moura <joao@crewai.com>
* docs: Update documentation for inject_date feature
Co-Authored-By: Joe Moura <joao@crewai.com>
* unnecesary
* new tests
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Joe Moura <joao@crewai.com>
Co-authored-by: João Moura <joaomdmoura@gmail.com>
- Add `HallucinationGuardrail` class as enterprise feature placeholder
- Update LLM guardrail events to support `HallucinationGuardrail` instances
- Add comprehensive tests for `HallucinationGuardrail` initialization and behavior
- Add integration tests for `HallucinationGuardrail` with task execution system
- Ensure no-op behavior always returns True
* Refactor Crew class memory initialization and enhance event handling
- Simplified the initialization of the external memory attribute in the Crew class.
- Updated memory system retrieval logic for consistency in key usage.
- Introduced a singleton pattern for the Telemetry class to ensure a single instance.
- Replaced telemetry usage in CrewEvaluator with event bus emissions for test results.
- Added new CrewTestResultEvent to handle crew test results more effectively.
- Updated event listener to process CrewTestResultEvent and log telemetry data accordingly.
- Enhanced tests to validate the singleton pattern in Telemetry and the new event handling logic.
* linted
* Remove unused telemetry attribute from Crew class memory initialization
* fix ordering of test
* Implement thread-safe singleton pattern in Telemetry class
- Introduced a threading lock to ensure safe instantiation of the Telemetry singleton.
- Updated the __new__ method to utilize double-checked locking for instance creation.
* Add reasoning attribute to Agent class
Co-Authored-By: Joe Moura <joao@crewai.com>
* Address PR feedback: improve type hints, error handling, refactor reasoning handler, and enhance tests and docs
Co-Authored-By: Joe Moura <joao@crewai.com>
* Implement function calling for reasoning and move prompts to translations
Co-Authored-By: Joe Moura <joao@crewai.com>
* Simplify function calling implementation with better error handling
Co-Authored-By: Joe Moura <joao@crewai.com>
* Enhance system prompts to leverage agent context (role, goal, backstory)
Co-Authored-By: Joe Moura <joao@crewai.com>
* Fix lint and type-checker issues
Co-Authored-By: Joe Moura <joao@crewai.com>
* Enhance system prompts to better leverage agent context
Co-Authored-By: Joe Moura <joao@crewai.com>
* Fix backstory access in reasoning handler for Python 3.12 compatibility
Co-Authored-By: Joe Moura <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Joe Moura <joao@crewai.com>
Co-authored-by: João Moura <joaomdmoura@gmail.com>
* Add markdown attribute to Task class for formatting responses in Markdown
Co-Authored-By: Joe Moura <joao@crewai.com>
* Enhance markdown feature based on PR feedback
Co-Authored-By: Joe Moura <joao@crewai.com>
* Fix lint error and validation error in test_markdown_task.py
Co-Authored-By: Joe Moura <joao@crewai.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Joe Moura <joao@crewai.com>
* Changed test case
* Addd new interaction with llama
* fixed linting issue
* Gemma Flaky test case
* Gemma Flaky test case
* Minor change
* Minor change
* Dropped API key
* Removed falky test case check file
