[dotnet-port] Port TodoList_Complete reason requirement from .NET

[qmuntal]

…#5902

Port the upstream .NET change that replaces the plain-ID slice in TodoList_Complete with a structured CompleteInput{ID, Reason} slice. Agents are now prompted to include a completion reason, which improves auditability of todo lifecycle events.

Changes:

• Add CompleteInput struct (id + reason fields)
• Update completeTool to accept []CompleteInput instead of []int
• Update tool description to mention the reason field
• Update default instructions to ask agents to include a reason
• Update all existing tests to use the new input format
• Add TestCompleteTodos_WithReason and TestCompleteToolDescription_MentionsReason

[Copilot]

Pull request overview

Ports the upstream .NET change to the todo harness so TodoList_Complete takes structured completion entries (ID + reason), prompting agents to include a completion reason for better auditability.

Changes:

• Introduces CompleteInput { id, reason } and updates TodoList_Complete to accept []CompleteInput.
• Updates default instructions and tool description to explicitly request a completion reason.
• Updates existing tests to use the new input shape and adds tests for the updated description / acceptance.

Show a summary per file

File	Description
agent/harness/todo/todo.go	Adds CompleteInput and updates TodoList_Complete schema/description/instructions to include a completion reason.
agent/harness/todo/todo_test.go	Migrates tool-call JSON payloads to the new {id, reason} format and adds reason/description-focused tests.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 2/2 changed files
• Comments generated: 2

[qmuntal]

@copilot fix codereview feedback,

[Copilot]

@copilot fix codereview feedback,

Done in the latest commit. Both pieces of feedback are addressed:

1. The TodoList_Complete handler now validates each entry upfront — strings.TrimSpace(item.Reason) == "" returns an error before any state mutation, so callers cannot slip past an empty or whitespace-only reason.
2. TestCompleteTodos_EmptyReasonIsRejected covers the empty-string and whitespace-only cases and asserts both that an error is returned and that the item is not marked complete.

[github-actions]

Generated by Go API Consistency Review Agent for issue #199 · ● 3.2M

[qmuntal]

@copilot fix #199 (comment)

[Copilot]

@copilot fix #199 (comment)

Removed the strict TrimSpace validation and the corresponding negative test to align with the upstream .NET TodoProvider, which accepts an empty/omitted reason without error. Added TestCompleteTodos_EmptyReasonIsAccepted to explicitly document this intentional parity with .NET.

[qmuntal]

@copilot fix golangci-lint

[Copilot]

@copilot fix golangci-lint

Fixed — the gofumpt violation was extra alignment spaces in the struct literal in TestCompleteTodos_EmptyReasonIsAccepted. Reformatted with gofumpt.

[github-actions]

Cross-repo Parity Review ✅

This PR faithfully ports the upstream .NET change (#5902) and stays aligned with the upstream implementation.

Checked against microsoft/agent-framework:

Surface	.NET (dotnet/src/Microsoft.Agents.AI/Harness/Todo/)	Go (agent/harness/todo/todo.go)
Input type name	TodoCompleteInput (internal)	CompleteInput (exported, todo.CompleteInput)
JSON field names	"id", "reason"	"id", "reason" ✅
Tool description	"Each entry has an ID and a reason describing how/why the item was completed."	Identical wording ✅
Default instructions	"Include a reason describing how the items were completed."	Identical wording ✅

Python: No TodoProvider equivalent exists in the upstream Python packages — no Python parity gap.

Note on visibility: The .NET type is internal, while the Go type is exported. This is intentional — Go's reflection-based tool argument binding requires exported types for correct JSON deserialization. No parity concern here.

No cross-repo consistency issues found.

Generated by Go API Consistency Review Agent for issue #199 · ● 9.2M · ◷