Matt Vanderpol
Summary
The Marker.io MCP server currently supports writing assignees via
issue_update_assignee
, but the read endpoints (
issues_list
and
issue_get
) don't return assignee data. Only
reporter
is exposed.
This makes a common workflow impossible from an AI assistant: "show me the tickets assigned to me."
Current behavior
-
issues_list
fields
enum:
id, markerId, title, description, status, priority, url, reporter, createdAt, updatedAt
-
issue_get
fields
enum: same as above plus
platform, comments, context, attachments
- No
assignee
option in either.
Requested behavior
1. Add
assignee
to the
fields
enum on both
issues_list
and
issue_get
. Shape should mirror
reporter
—
{ id, name, email }
(or array of those, if multi-assignee is supported).
2. Add an
assigneeId
filter parameter to
issues_list
so clients can fetch "issues assigned to user X" in a single call rather than paging the whole project and filtering client-side.
Why this matters
- The write side already exists (
issue_update_assignee
), so the data model supports it — this is an asymmetry between read and write surfaces.
- For agent / MCP use, "what's on my plate" is the single most common query against an issue tracker. Without
assignee
on read, every assistant has to ask the user to manually identify their own tickets.