fix: restore thread id in get_review_comments response

[omid-ant]

Summary

Restore the thread id field to MinimalReviewThread so that pull_request_review_write with method: resolve_thread / unresolve_thread is actually usable.

Why

#2062 dropped id from the minimized review-thread output, noting it was "not used for any subsequent tool call". While this was true at the time, #1919 later added resolve_thread / unresolve_thread, both of which require a thread node ID. The tool's own docstring reads:

Get thread IDs from pull_request_read with method get_review_comments.

…but that tool no longer returns them, so there's no way to resolve a thread without leaving the MCP server. This results in callers of this tool being unable to resolve threads through the resolve_thread tool.

The GraphQL query (reviewThreadNode in pkg/github/pullrequests.go) already fetches ID — it's just dropped during minimization.

Also see: #2229

What changed

• Add ID string to MinimalReviewThread
• Populate it in convertToMinimalReviewThread via fmt.Sprintf("%v", thread.ID)

MCP impact

• No tool or API changes
• Tool schema or behavior changed — pull_request_read (get_review_comments) now includes id per thread
• New tool added

Prompts tested (tool changes only)

• "List the review threads on PR [docker] build arm64 #42 in owner/repo, then resolve the first one"

Security / limits

• No security or limits impact
• Auth / permissions considered
• Data exposure, filtering, or token/size limits considered — ~40 chars/thread ("id":"PRRT_kwDO…",), negligible impact on Reduce context usage for get_pull_request_review_comments using pull_request_read #2062's 14.5% savings

Tool renaming

• I am renaming tools as part of this PR (e.g. a part of a consolidation effort)
  • I have added the new tool aliases in deprecated_tool_aliases.go
• I am not renaming tools as part of this PR

Docs

• Not needed
• Updated (README / docs / examples)

Fixes #2331

[MitchTalmadge]

This is a blocker to our workflow. Currently we are unable to resolve threads automatically through MCP using Claude, for example, because it does not know the IDs of the threads. Excited to see this PR and hope it can be approved soon.

[SamMorrowDrums]

Tracks #2331.

[scottlz0310-user]

Supporting this PR. This is exactly the gap that #2229 identified —
the id field being absent made resolve_thread unreachable from
an MCP-only session. Would be great to see this merged.

[scottlz0310-user]

As the reporter of #2229, I checked out this branch, built it locally, and verified the
behavior.

Verified:

• The id field now correctly appears in the JSON response from get_review_comments
  (e.g., "id": "PRRT_kwDO...")
• All existing tests pass (go test ./...)

One suggestion: the existing test for get_review_comments already provides "id" in the
mock data but never asserts thread.ID in the validation. Adding an assertion like
assert.Equal(t, "RT_kwDOA0xdyM4AX1Yz", thread.ID) would prevent silent regressions in
the future.