Feature: Make trace_include_sensitive_data configurable via environment variable

[Kunmeer-SyedMohamedHyder]

Description:

This PR adds support for configuring trace_include_sensitive_data in RunConfig using an environment variable: OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA.

Previously, this field always defaulted to True, and there was no way to control it without explicitly passing it into RunConfig. With this change, if the field is not set explicitly, its value will be pulled from the environment. This provides more flexibility, especially in environments where trace sensitivity and data privacy are important.

The environment variable supports common boolean values like true, false, 1, 0, yes, and no (case-insensitive).

Fixes #1183

Behavior Summary:

• If trace_include_sensitive_data is passed explicitly, that value is used.
• Otherwise, the value defaults based on the OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA environment variable.
• If the environment variable is not set, it falls back to True (preserving current behavior).

Example usage:

export OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA=false

config = RunConfig()  # trace_include_sensitive_data will be False

config = RunConfig(trace_include_sensitive_data=True)  # Explicit override, will be True

[github-actions]

This PR is stale because it has been open for 10 days with no activity.

[loiccordone]

Thank you for the PR @Kunmeer-SyedMohamedHyder , any blocking points before approving it @seratch?
Thanks all

[Copilot]

Pull Request Overview

This PR adds environment variable configuration for the trace_include_sensitive_data field in RunConfig. Previously, this field was hardcoded to True, but now it can be controlled via the OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA environment variable while maintaining backward compatibility.

• Introduces a new helper function _default_trace_include_sensitive_data() to parse the environment variable
• Changes trace_include_sensitive_data from a simple default value to use field(default_factory=...)
• Adds comprehensive test coverage for various boolean string formats and precedence behavior

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/agents/run.py	Implements environment variable parsing logic and updates RunConfig field definition
tests/test_run_config.py	Adds test cases for environment variable parsing, default behavior, and explicit override precedence

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The boolean parsing logic is incomplete and will incorrectly handle invalid values. For example, setting the environment variable to "invalid" will return False, but it should either raise an error or have documented fallback behavior. Consider adding validation or explicit handling for unrecognized values.

Suggested change

	return val.strip().lower() in ("1", "true", "yes", "on")
	val = os.getenv("OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA", "true").strip().lower()
	truthy = ("1", "true", "yes", "on")
	falsy = ("0", "false", "no", "off")
	if val in truthy:
	return True
	elif val in falsy:
	return False
	else:
	raise ValueError(
	f"Invalid value for OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA: '{val}'. "
	f"Expected one of {truthy + falsy}."
	)

[seratch]

I think having this enhancement is good; @rm-openai any thoughts?