[SergeyMenshykh]

Motivation & Context

Make MCP skill archive loading more predictable across Python and .NET, with clearer diagnostics and expanded regression coverage.

Description & Review Guide

• What are the major changes? Refine internal archive loading, integrate warning diagnostics, update related documentation, and extend tests in both implementations.
• What is the impact of these changes? More predictable archive handling without new public APIs or configuration requirements.
• What do you want reviewers to focus on? Implementation consistency, resource lookup behavior, diagnostics, and regression coverage.

Related Issue

No public issue linked.

Contribution Checklist

• The code builds clean without any errors or warnings
• All unit tests pass, and I have added new tests where possible (focused .NET and Python suites pass; the full suite has not been run locally).
• The PR follows the Contribution Guidelines
• This PR is linked to an issue and there is no other open PR for this issue (see Related Issue above).
• This is not a breaking change. If it is a breaking change, add the breaking change label (or add "[BREAKING]" to the title prefix, before or after any language prefix) — a workflow keeps the label and the title prefix in sync automatically.

[Copilot]

Copilot review overview

🟢 Approval recommended

The implementations consistently preserve first-seen content, emit diagnostics, retain safety limits, and include comprehensive regression coverage.

Review effort: Balanced
Findings: None

What changed in this PR

Improves MCP skill archive collision handling across Python and .NET.

Changes:

• Preserves the first colliding archive member and logs warnings.
• Prevents duplicate decompression and file overwrites.
• Adds extensive collision and limit regression coverage.

File	Description
python/​packages/​core/​agent_framework/​_skills.py	Adds case-insensitive in-memory collision handling.
python/​packages/​core/​tests/​core/​test_mcp_skills.py	Tests collision aliases, ordering, warnings, and limits.
dotnet/​src/​Microsoft.Agents.AI.Mcp/​Skills/​Loaders/​AgentMcpSkillArchiveExtractor.cs	Adds collision-safe extraction and diagnostics.
dotnet/​src/​Microsoft.Agents.AI.Mcp/​Skills/​Loaders/​ArchiveEntryLoader.cs	Passes logging into extraction and reuses path sanitization.
dotnet/​src/​Microsoft.Agents.AI.Mcp/​Skills/​AgentMcpSkillsSource.cs	Documents archive collision behavior.
dotnet/​tests/​Microsoft.Agents.AI.Mcp.UnitTests/​Skills/​AgentMcpSkillsSourceArchiveTests.cs	Adds archive collision regression tests.

---

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

[github-actions]

MAF Automated Review — Iteration 1

Result: No findings
Scope: full PR (1 commit(s)): bea7b7d4af85
Model: gpt-5.6-sol

Overview

The change makes duplicate archive handling first-member-wins, reports collisions without exposing file contents, and preserves per-runtime resource lookup semantics. Path normalization, extraction limits, digest checks, pre-decompression duplicate detection, and .NET partial-extraction cleanup prevent the examined collision cases from producing a publishable defect.

Reviewed the supplied pull-request change set across correctness, security/reliability, architecture, and failure behavior.
No publishable findings remained after source verification for this scope.