taylorwilsdon/google_workspace_mcp v1.28.0

on GitHub

This release adds named range management to Google Sheets and makes Gmail more honest about what it reads and writes: forwarded and wrapped emails are now visible, bad thread IDs fail loudly, and label edits stop resetting visibility. It also hardens Office file text extraction against files that expand far beyond their download size.

New

• gsheets: New manage_named_range tool lists, creates, renames, re-targets and deletes named ranges using A1 notation (complete tier). feat: manage gSheets “Named Range” operations: list, create, update and delete by @antoine2711 in #1146

Gmail fixes

• gmail: Updating a label (e.g. a rename) no longer resets hidden labels to visible, because it now keeps the label's stored visibility unless you change it. fix(gmail): keep a label's visibility when updating it by @ConnorMoss02 in #1095
• gmail: Message and thread reads now show emails wrapped inside another email (forward-as-attachment, bounces, digests, moderation notices) with their headers and body, and label their files [in attached message]. fix(gmail): show emails wrapped inside another email (message/rfc822) by @123andy in #1151
• gmail: draft_gmail_message and send_gmail_message now return a clear error when Gmail rejects a thread_id, such as a token copied from a Gmail web URL. Before, the draft tool created a standalone draft outside the thread and still reported success. fix(gmail): reject an unresolvable thread_id instead of creating an orphan draft by @123andy in #1150

Office extraction hardening

• core: mc:Choice elements that share a scope now reuse a single prefix map, so crafted Office XML can no longer inflate parser memory to hundreds of times its size. fix(core): share the prefix map between mc:Choice elements in one scope by @123andy in #1154
• core: Text extraction from .docx/.xlsx/.pptx now stops at a per-file expansion budget (new WORKSPACE_MCP_MAX_OFFICE_XML_BYTES, default 25 MiB, 0 to disable), so a small ZIP bomb can't exhaust server memory. get_drive_file_content and get_doc_content report files over the limit as too large instead of as damaged. fix(core): bound how far an Office file may expand during text extraction by @123andy in #1153

New Contributors

• @antoine2711 made their first contribution in #1146

Full Changelog: v1.27.1...v1.28.0