[Feature Request] End-to-end testing for OpenAI Agents SDK examples

[jssmith]

Is your feature request related to a problem? Please describe.

We need to add end-to-end tests to OpenAI Agents SDK samples to ensure that they remain stable.

Describe the solution you'd like

• Using the commands in the README, run each sample by launching the worker and the workflow runner program
• In some cases, additional resources, such as MCP servers or Postgres databases, may also need to run
• Some samples are interactive. Support this by sending commands to standard in.
• The core functionality should be conventional programmatic code.
• If the workflow runner exits with success, consider the run successful.
• Clean up all resources.

Additional context

• Optionally, verify: 1/ commands run match the README files (LLM or traditional search), 2/ Outputs match expected outputs (using LLMs).

[cretz]

These tests should also serve as samples for how users might write their tests in the same situations