How to read Jira issue comments with atlassian-mcp-server getJiraIssue?
Using atlassian-mcp-server, the getJiraIssue tool returns only basic fields (summary, description, status, assignee) and omits comments, while addCommentToJiraIssue can write comments but not read them, and there is no getJiraIssueComments tool. How can I retrieve existing Jira issue comments through MCP without switching to a browser? The thread shows a fields-parameter workaround and a maintainer confirmation, but no client or OS details were given. The thread shows a fields-parameter workaround and a maintainer confirmation.
Originally asked on GitHub: https://github.com/atlassian/atlassian-mcp-server/issues/88
4 answers
The feature request was closed as already-supported: no separate
getJiraIssueCommentstool was added. Reading Jira comments works through the existinggetJiraIssuetool by adding"comment"to thefieldsarray. A maintainer confirmed this is documented, supported behavior.Call
getJiraIssuelike this:jsonCallMcpTool: getJiraIssue Arguments: { "cloudId": "YOUR_CLOUD_ID", "issueIdOrKey": "TICKET-123", "fields": ["comment", "summary", "description"], "responseContentFormat": "markdown" }The response includes the comments under
fields.comment.commentsplus a convenience top-levelcommentsarray. Each comment contains the body, author (account ID and display name), created and updated timestamps, and visibility. The body is converted to the format you set inresponseContentFormat("markdown"by default, or"adf"), so you do not have to parse raw ADF yourself when using markdown.A user also confirmed the same approach works with a broader fields list:
["comment", "comments", "summary", "description", "*all"].One limitation the maintainer noted: if you need to paginate through a very large comment thread on its own without pulling the rest of the issue payload,
getJiraIssuedoes not handle that. File a separate issue if that is your workflow.Did this work for you?What @jatinkrmalik tried on GitHub (Worked)
A maintainer explained that passing "comment" in the fields array of getJiraIssue is supported: comments come back under fields.comment.comments and as a top-level comments array, with bodies converted per responseContentFormat ("markdown" by default).
Did this work for you?What @stu4g tried on GitHub (Worked)
A user worked around the missing read-comments tool by calling getJiraIssue with fields ["comment", "comments", "summary", "description", "*all"] and responseContentFormat "markdown" to force comment retrieval.
Did this work for you?What a GitHub user tried on GitHub (Didn't work)
The reporter tried getJiraIssue and fetch with a Jira issue ARI; both return only basic fields without comments, and addCommentToJiraIssue only writes comments rather than reading them.
Source: https://github.com/atlassian/atlassian-mcp-server/issues/88 (opens external site)
Did this work for you?
Related questions
- Atlassian MCP server broken after Claude Code dropped SSE: what works?
- Atlassian MCP server in Claude Code/Cursor: "We are having trouble completing this action" error
- How to link Jira issues (issuelinks) with atlassian-mcp-server MCP tools?
- Atlassian MCP server drops connection after 15-20 min with 401 Unauthorized
- atlassian-mcp-server: Can I authenticate with an API token instead of OAuth?