Masheev Developers

Open Markdown · Agent index

Jira issues

Create or link a Jira issue from a conversation, then follow it until it is fixed. Status changes and comments come back to the conversation as internal notes for your team. Masheev works with Jira Cloud.

Jira issue linking is in beta. See Availability.

Connect Jira

Jira uses the same Atlassian connection as Confluence, with a separate approval.

  1. An admin opens Settings → Integrations.
  2. If Atlassian is not connected yet, connect it and choose your site.
  3. Enable Jira on the Atlassian connection. Atlassian asks you to approve read and write access to Jira work items. Masheev asks for this only when you enable Jira.
  4. Pick the projects your team may file issues into, and a default issue type.

Teammates can only create or link issues in the projects you pick. You can change the list at any time.

Create an issue from a conversation

  1. In a conversation, open Create Jira issue.
  2. Masheev drafts the issue from the conversation: a summary, steps to reproduce, expected result, actual result and impact.
  3. Before you file, Masheev shows open issues in your projects that look similar. Link one of those instead if it is the same problem.
  4. Review and edit the draft, pick the project and issue type, and fill in any fields your Jira project requires.
  5. Create the issue. It is linked to the conversation.

The draft is only a suggestion. Nothing is filed until a teammate creates it.

What goes into the issue

  • Email addresses and phone numbers are removed from the summary and description, including anything you typed.
  • The description ends with an opaque customer reference, for example Customer reference: cust-3f9a1c2b7e4d. It is random and not derived from the customer's details. Use the conversation in Masheev to see who reported it.
  • The issue is created with the Atlassian account that enabled Jira, so it appears in Jira as filed by that person.

Creating the same draft twice does not create two issues. If a request is retried, Masheev returns the issue it already filed.

Enter an issue key such as SUP-142 to follow an issue that already exists. The issue must be in one of your enabled projects. Comments posted before you link the issue are not imported.

Follow the issue

The conversation shows each linked issue with its current status. About every 30 minutes Masheev checks linked issues and:

  • updates the status,
  • adds each new Jira comment to the conversation as an internal note,
  • when the issue moves to a done status, adds an internal note with a suggested reply you can send, for example: "Good news: the issue you reported has been fixed. Thanks for your patience, and let us know if you still see it."

After the done note, Masheev stops following the issue. Unlinking an issue stops following it and leaves the issue in Jira unchanged.

Procedure reference

Call these through the API or MCP. Hosted MCP tool names use __ in place of the dot, for example tickets__create.

ProcedureWhoWhat it does
integrationStatus.setJiraSettingsAdminSave the allowed projects (up to 20) and default issue type.
integrationStatus.selectAtlassianSiteAdminChoose the Atlassian site when your account has several.
tickets.configMemberWhether Jira is ready, the site name and allowed projects.
tickets.projectsMemberProjects the connection can see.
tickets.createMetaMemberIssue types and required fields for an allowed project.
tickets.draftMemberAI draft from a conversation. Files nothing.
tickets.similarMemberOpen issues in allowed projects that match a short description.
tickets.createMemberCreate an issue from a reviewed draft and link it.
tickets.linkMemberFollow an existing issue from a conversation.
tickets.listMemberIssues a conversation follows, with status.
tickets.unlinkMemberStop following an issue.

Create an issue

draftId is any unique id your client generates for this draft, 8 to 64 characters. Reuse it when you retry so the issue is created once.

{
  "orgId": "YOUR_WORKSPACE_ID",
  "conversationId": "CONVERSATION_ID",
  "projectKey": "SUP",
  "issueTypeId": "10004",
  "summary": "Checkout button does nothing on Safari",
  "description": "Steps to reproduce:\n1. Add an item\n2. Open checkout on Safari 18\n\nExpected:\nPayment form opens\n\nActual:\nNothing happens\n\nImpact:\nSafari customers cannot pay",
  "draftId": "d7c1e0a2-5b7f-4c1e"
}

Pass required custom fields from tickets.createMeta in fields, keyed by field key.

Save Jira settings

{
  "projects": ["SUP", "WEB"],
  "issueTypeId": "10004"
}

Rate limits apply per workspace: tickets.draft 30 per 10 minutes, tickets.create 30 per hour.

What Masheev never does

  • It never messages a customer about an issue. The suggested reply is a note for your team to review and send.
  • Customers cannot look up issues, and AI agents have no Jira tools.
  • It never files an issue on its own. A teammate always creates it.
  • It never writes customer email addresses or phone numbers into Jira.
  • It never files into projects you did not enable.

Disconnecting Atlassian stops issue syncing. Issues already in Jira stay there.

Privacy

Masheev stores the issue key, title, status and link for each linked issue, and adds new comments to the conversation as internal notes. It does not store Atlassian account IDs, names, email addresses or avatars; a mention in a comment shows as @teammate. See the Privacy Policy (section 6.5) and the DPA.