Rebar Wiki

jira-integration

3 min read

Jira Integration

tools jira mcp project-management

Connect Rebar to Jira for automatic ticket context in expertise.yaml. Uses the MCP Jira server so Claude can read sprints, tickets, and blockers directly.

Setup

1. Install the MCP Jira Server

Add the Atlassian MCP server to your Claude Desktop config (claude_desktop_config.json):

{
  "mcpServers": {
    "jira": {
      "command": "npx",
      "args": [
        "-y",
        "@anthropic/mcp-remote",
        "https://mcp.atlassian.com/v1/sse"
      ]
    }
  }
}

Restart Claude Desktop. It will prompt for Atlassian OAuth on first use.

For Claude Code, add to .mcp.json in your project root:

{
  "mcpServers": {
    "jira": {
      "command": "npx",
      "args": [
        "-y",
        "@anthropic/mcp-remote",
        "https://mcp.atlassian.com/v1/sse"
      ]
    }
  }
}

2. Verify Access

After auth, test with a simple query:

> Search for open issues in project DEMO

Found 14 open issues in DEMO:
  DEMO-456: Rollback endpoint returns 500 on concurrent requests (Bug, High)
  DEMO-455: Add deployment metrics dashboard (Story, Medium)
  ...

What /discover Pulls from Jira

When you run /discover {client}, Rebar reads Jira to seed expertise.yaml with project state:

DataWhere It GoesExample
Open ticket count by typeproject_health.backlogbugs: 3, stories: 8, tasks: 2
Current sprint ticketsproject_health.current_sprintsprint: "Sprint 14", committed: 8, completed: 3
Blockers and dependenciesproject_health.blockers"DEMO-456 blocked by DEMO-123 (auth service migration)"
Recent velocityproject_health.velocitylast_3_sprints: [18, 21, 16]
Component ownersteam.ownership"auth-service: @sarah, deploy-api: @brian"

Example expertise.yaml section

After /discover with Jira connected:

project_health:
  backlog:
    total: 23
    bugs: 3
    stories: 14
    tasks: 6
  current_sprint:
    name: "Sprint 14"
    goal: "Deployment metrics + rollback hardening"
    committed: 8
    completed: 3
    days_remaining: 5
  blockers:
    - ticket: DEMO-456
      title: "Rollback endpoint returns 500 on concurrent requests"
      blocked_by: DEMO-123
      reason: "Auth service migration must complete before rollback can verify service tokens"
      days_blocked: 3
  velocity:
    last_3_sprints: [18, 21, 16]
    average: 18.3

What /meeting Extracts

When meeting notes reference Jira tickets, Rebar links them:

Meeting notes mention: "DEMO-456 is blocked by the auth migration"

/meeting extracts:
  - Links DEMO-456 to DEMO-123 dependency
  - Captures "3 days blocked" timeline
  - Notes owner (@sarah) committed to unblocking by Thursday
  - Appends to unvalidated_observations:
    "DEMO-456 blocked by DEMO-123 (auth migration) -- @sarah targeting Thursday unblock"

Jira Queries Rebar Uses

These JQL queries power the integration:

# Open issues for a project
project = DEMO AND status != Done ORDER BY priority DESC

# Current sprint
project = DEMO AND sprint in openSprints()

# Blockers
project = DEMO AND status = "Blocked" OR issueFunction in linkedIssuesOf("is blocked by")

# Recent completions (velocity calc)
project = DEMO AND status = Done AND resolved >= -21d

# Issues by component/label
project = DEMO AND component = "auth-service"

Tips

  • Run /discover at the start of each sprint to refresh project health data
  • Blocker information is especially valuable for /brief — it surfaces what’s stuck
  • Jira ticket IDs in expertise.yaml observations make them searchable
  • The MCP server respects Jira permissions — Claude only sees what your account can see

Source: MCP Atlassian server docs, tested with Jira Cloud

Graph View

Backlinks