Docs · Features
Worklist
My work is the cross-project reviewer worklist: one page listing the paused review steps that name you, the queries assigned to you, and the documents in your scope needing quality attention. It is a page you visit, not a notification center, and it never assigns, gates, or moves anything.
What it does#
My work answers one question across every project you can see: what is waiting on me? It merges three kinds of item, newest first. The Review section lists paused workflow steps whose step definition names your email. The Query section lists threads assigned to you that are open or answered, plus your own threads that were just answered, so you can act on the reply; it disappears entirely when queries are turned off, and the other two sections are unaffected. The Quality section lists every document you can see with at least one live major or critical LQA finding within a recency window, ninety days by default and adjustable per request; it is reviewer-role or higher, grouped by document and locale so one noisy document is one row, and it is a derived list rather than an assignment, since nothing in Localization OS assigns an LQA task to a specific person.
The worklist reads only: opening it never changes anything, assigns anything, or gates a workflow. There is also an off-by-default view that swaps the Quality section for the raw list of segments a quality gate routed to a human for review, useful occasionally but noisier than the default findings-based view on an established install.
Getting started#
Open My work from the navigation to see everything waiting on you across every project you can see. Narrow the quality window, or switch to the raw needs-review view, from the same page. From the CLI, --actor is a global flag and belongs before the subcommand, and it is required for an in-process work list.
loc-tms --actor you@example.com work list
loc-tms --actor you@example.com work list --days 30
loc-tms --actor you@example.com work list --quality needs_reviewThe same page, with the same filters, is available over the API for building an integration, and as an MCP tool for an agent working on your behalf.
Works with#
- Queries: the Query section lists the threads assigned to you and your own answered ones; turning queries off removes only this section.
- Workflows: the Review section lists paused human-review steps whose definition names you, and each row links to the run where you post-edit and resume it.
- Reports: the Quality section reads the same major and critical LQA findings the QA report measures; the worklist is the per-person view of that same data, not a separate measurement.
Current limits#
Review assignment lives on the workflow's step definition, not on an individual run. If a review step names you, every paused run of that workflow shows up, since per-run assignment does not exist yet.
The Quality section is a derived list, not an assignment. It is every document you can see with open major or critical findings, not a signal that anyone was specifically told to fix them, and a withdrawn finding never appears there.
It is a page you visit, not a notification center. There is no unread count and no digest; the only push notification anywhere in the product is the email sent for a query assignment.
Everything shown is scoped to what you can already see. A customer-bound account sees only that customer's work, and the total count shown only ever reflects in-scope rows.
The API needs a per-user key to know whose work to show: a shared connector token has no individual user behind it and is refused with a named message rather than returning an empty list that would read as "you have no work."