This pull request introduces functionality for visualizing agent
structures using Graphviz. The changes include adding a new dependency,
implementing functions to generate and draw graphs, and adding tests for
these functions.
New functionality for visualizing agent structures:
* Added `graphviz` as a new dependency in `pyproject.toml`.
* Implemented functions in `src/agents/visualizations.py` to generate
and draw graphs for agents using Graphviz. These functions include
`get_main_graph`, `get_all_nodes`, `get_all_edges`, and `draw_graph`.
Testing the new visualization functionality:
* Added tests in `tests/test_visualizations.py` to verify the
correctness of the graph generation and drawing functions. The tests
cover `get_main_graph`, `get_all_nodes`, `get_all_edges`, and
`draw_graph`.
For example, given the following code:
```python
from agents import Agent, function_tool
from agents.visualizations import draw_graph
@function_tool
def get_weather(city: str) -> str:
return f"The weather in {city} is sunny."
spanish_agent = Agent(
name="Spanish agent",
instructions="You only speak Spanish.",
)
english_agent = Agent(
name="English agent",
instructions="You only speak English",
)
triage_agent = Agent(
name="Triage agent",
instructions="Handoff to the appropriate agent based on the language of the request.",
handoffs=[spanish_agent, english_agent],
tools=[get_weather],
)
draw_graph(triage_agent)
```
Generates the following image:
<img width="614" alt="Screenshot 2025-03-13 at 18 36 23"
src="https://github.com/user-attachments/assets/d01fe502-6886-4efb-aaf8-c92e4524b0fe"
/>
## Summary:
#263 added this behavior. The goal was to prevent infinite loops when tool choice was set. The key change I'm making is:
1. Making it configurable on the agent.
2. Doing bookkeeping in the Runner to track this, to prevent mutating agents.
3. Not resetting the global tool choice in RunConfig.
## Test Plan:
Unit tests.
.
### Summary:
This enables users to **use** MCP inside the SDK.
1. You add a list of MCP servers to `Agent`, via `mcp_server=[...]`
2. When an agent runs, we look up its MCP tools and add them to the list
of tools.
3. When a tool call occurs, we call the relevant MCP server.
Notes:
1. There's some refactoring to make sure we send the full list of tools
to the Runner/Model etc.
2. Right now, you could have a locally defined tool that conflicts with
an MCP defined tool. I didn't add errors for that, will do in a
followup.
### Test Plan:
See unit tests. Also has an end to end example next PR.
---
[//]: # (BEGIN SAPLING FOOTER)
* #324
* #322
* __->__ #321
* #320
# Fix potential infinite tool call loop by resetting tool_choice after
tool execution
## Summary
This PR fixes an issue where setting `tool_choice` to "required" or a
specific function name could cause models to get stuck in an infinite
tool call loop.
When `tool_choice` is set to force tool usage, this setting persists
across model invocations. This PR automatically resets `tool_choice` to
"auto" after tool execution, allowing the model to decide whether to
make additional tool calls in subsequent turns.
Unlike using `tool_use_behavior="stop_on_first_tool"`, this approach
lets the model continue processing tool results while preventing forced
repeated tool calls.
## Test plan
- Added tests to verify tool_choice reset behavior for both agent and
run_config settings
- Added integration test to verify the solution prevents infinite loops
- All tests pass
## Checks
- [x] I've added new tests for the fix
- [x] I've updated the relevant documentation (added comment in code)
- [x] I've run `make lint` and `make format`
- [x] I've made sure tests pass
### Summary:
This enables users to **use** MCP inside the SDK.
1. You add a list of MCP servers to `Agent`, via `mcp_server=[...]`
2. When an agent runs, we look up its MCP tools and add them to the list of tools.
3. When a tool call occurs, we call the relevant MCP server.
Notes:
1. There's some refactoring to make sure we send the full list of tools to the Runner/Model etc.
2. Right now, you could have a locally defined tool that conflicts with an MCP defined tool. I didn't add errors for that, will do in a followup.
### Test Plan:
See unit tests. Also has an end to end example next PR.
This fixes the issue where the original agent's model_settings was being directly modified during the tool choice reset process. The original implementation caused the agent's tool_choice to unintentionally reset to "auto" for subsequent runs, which could be unexpected behavior.
The fix creates new copies of the agent and model settings objects using dataclasses.replace() instead of modifying the original objects. This ensures that the tool choice reset is limited to the current run only, maintaining the expected behavior for sequential runs with the same agent.
Addresses feedback from @baderalfahad about the agent instance being modified when it should maintain its original state between runs.
- Refactor tool_choice reset to target only problematic edge cases
- Replace manual ModelSettings recreation with dataclasses.replace
- Fix line length and error handling lint issues in tests
This PR introduces a `strict_mode: bool = True` option to
`@function_tool`, allowing optional parameters when set to False. This
change enables more flexibility while maintaining strict JSON schema
validation by default.
resolves#43
## Changes:
- Added `strict_mode` parameter to `@function_tool` and passed it to
`function_schema` and `FunctionTool`.
- Updated `function_schema.py` to respect `strict_mode` and allow
optional parameters when set to False.
- Added unit tests to verify optional parameters work correctly,
including multiple optional params with different types.
## Tests:
- Verified function calls with missing optional parameters behave as
expected.
- Added async tests to validate behavior under different configurations.
- Enhanced the readability of the test case by reformatting the expected output and input message structures.
- This change maintains the same functionality while making the test code cleaner and easier to understand.
- The _Converter.items_to_messages method was incorrectly rejecting 'assistant'
as a valid role in conversation messages, causing runtime errors when processing
standard chat completion message formats.
- This fix enables proper handling of
complete conversation contexts that include both user and assistant messages.
Rohan Mehta
Martín Bravo
xianghuijin
Alex Hall
Dominik Kundel
Raduan77
MartinEBravo
Muhammad Junaid
Jai0401