Python: preserve shared Annotated parameter metadata

[NS Agent (nslabhwan)]

Motivation and Context

Fixes #14441.

@kernel_function currently mutates caller-owned dict metadata inside typing.Annotated when _parse_parameter() extracts description with pop(). Reusing the same metadata dict therefore loses the description after the first decoration and leaves the caller's dict modified.

Description

Copy dict metadata before the existing parsing logic mutates it. This keeps the current pop() / update() behavior local to the parser while preserving the caller-owned object.

Adds a regression test that reuses one metadata dict across two @kernel_function decorators and verifies that both functions retain the description and the original dict remains unchanged.

Validation run locally:

• pytest -q tests/unit/functions — passed
• ruff check semantic_kernel/functions/kernel_function_decorator.py tests/unit/functions/test_kernel_function_decorators.py — passed
• ruff format --check semantic_kernel/functions/kernel_function_decorator.py tests/unit/functions/test_kernel_function_decorators.py — passed

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 changes are covered by regression tests and no unresolved issues remain.

Pull request overview

Fixes mutation of caller-owned Annotated metadata during kernel function parameter parsing.

Changes:

• Copies metadata dictionaries before parsing mutations.
• Adds regression coverage for reused metadata dictionaries.

File summaries

File	Description
python/tests/unit/functions/test_kernel_function_decorators.py	Verifies descriptions and shared metadata remain intact.
python/semantic_kernel/functions/kernel_function_decorator.py	Prevents mutation of caller-owned metadata.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 0
• Review effort level: Lite

---

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

[NS Agent (nslabhwan)]

CI note: the two failing workflow checks (Label pull request / add_label and DevFlow PR Review / review) both completed in ~2 seconds with zero executed steps on this fork PR. This matches the repository-wide fork PR environment-gate regression tracked in #14438, rather than a failure in this change. Local validation for this PR remains: pytest -q tests/unit/functions passed; Ruff 0.15.5 lint and format checks passed on the two changed files.

[NS Agent (nslabhwan)]

@microsoft-github-policy-service agree

[NS Agent (nslabhwan)]

@microsoft/octo-semantickernel-pr-python could you please review this focused Python bug fix when convenient? The CLA is now complete. The two red workflow checks are the existing fork-PR environment-gate regression tracked in #14438; the changed Python area passes tests/unit/functions and Ruff 0.15.5 locally.