[Li-john1021]

Motivation and Context

KernelFunctionSelectionStrategy and KernelFunctionTerminationStrategy declare optional result_parser fields whose defaults are zero-argument lambdas, but select_agent / should_agent_terminate always call self.result_parser(result) with one argument. Using either strategy without a custom parser therefore always raises TypeError.

Fixes #14589

Description

Update the defaults so they accept the function result:

• Selection: lambda result: str(result).strip() — use the function output as the agent name (aligned with the .NET default that reads a string from the result).
• Termination: lambda result: str(result).strip().lower() == "true" — treat a case-insensitive "true" as terminate.

For termination, .NET’s default is (_) => true (always terminate). I followed the issue’s suggestion instead, because an always-true default would ignore the kernel function output and stop the chat on the first check. Happy to switch to lambda _: True for .NET parity if maintainers prefer that.

Also adds regression tests that construct both strategies without a custom result_parser.

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the SK Contribution Guidelines and the pre-submission formatting script raises no violations
• All unit tests pass, and I have added new tests where possible
• I didn't break anyone 😄

[Copilot]

🟢 Approval recommended

The defaults now match their call sites, and regression tests cover both strategies.

0 open findings

What changed in this PR

Fixes default parser arity and behavior for Python kernel-function agent strategies.

Changes:

• Parses function results into agent names or termination decisions.
• Adds regression tests for default parser behavior.

File	Description
test_kernel_function_termination_strategy.py	Tests true and false default parsing.
test_kernel_function_selection_strategy.py	Tests default agent-name parsing.
kernel_function_termination_strategy.py	Adds a one-argument boolean parser.
kernel_function_selection_strategy.py	Adds a one-argument agent-name parser.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Li-john1021]

Thanks for the approval and for queuing this, Eduard van Valkenburg (@eavanvalkenburg)!

It looks like the PR was removed from the merge queue because the merge-group Python Pre-Merge Integration Tests - Agents hit an Azure infra timeout (unrelated to this change):

test_file_search_invoke → azure.core.exceptions.DecodeError / vector_stores request timed out
(run 40325498196)

Unit tests and other checks on the PR itself are green. I don’t have permission to re-enable auto-merge / re-queue — would you mind adding it back to the merge queue when convenient? Happy to help if anything else is needed. Thank you!