Introduction

Agile methodologies are deeply rooted in enterprise daily workflows. Interacting with Jira issues for ticket management and Confluence for knowledge sharing and technical processes is an essential part of software development.

With the Model Context Protocol (MCP), the OpenCode CLI workflow begins directly inside your terminal by reading Jira issues, referencing linked Confluence pages for workflow instructions, and executing development tasks with full context of user stories and specifications.

Why This Matters

Switching context between code in the terminal and browser tabs with tickets and wikis slows down momentum and disrupts focus. Having full planning and deep understanding of a task—based directly on Jira issues and Confluence documentation—allows OpenCode to act as a truly context-aware coding assistant.

By integrating Atlassian tools via MCP, OpenCode can search for relevant tickets, read specification docs, follow organizational workflows, and keep project management tools updated without ever leaving the command line.

Prerequisites

  • OpenCode CLI installed and configured
  • Podman installed and running locally
  • An active Atlassian Cloud account (Jira & Confluence)
  • An Atlassian API Token generated for your user account

Step 1: Generate Atlassian API Token

To allow the MCP server to authenticate with your Atlassian Cloud instance, generate an API token:

  1. Log in to id.atlassian.com/manage-profile/security/api-tokens.
  2. Click Create API token.
  3. Give your token a label (e.g., opencode-mcp) and copy the generated token.

Step 2: Set Environment Variables

Export your Atlassian credentials in your shell environment:

export ATLASSIAN_EMAIL="your-email@example.com"
export ATLASSIAN_API_TOKEN="your-atlassian-api-token"

(Tip: Add these exports to your ~/.bashrc or shell profile to make them persistent).

Step 3: Configure OpenCode MCP Settings

Edit your OpenCode configuration file (~/.config/opencode/opencode.json) to register the Atlassian MCP server container:

mkdir -p ~/.config/opencode
nano ~/.config/opencode/opencode.json

Add the mcp server configuration to your opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "model": "google-vertex/gemini-3.6-flash",
  "providers": {
    "google-vertex": {
      "package": "@opencode-ai/ai/providers/google-vertex",
      "options": {
        "location": "us-central1"
      },
      "models": {
        "gemini-3.6-flash": {
          "name": "Gemini 3.6 Flash"
        }
      }
    }
  },
  "mcp": {
    "atlassian": {
      "type": "local",
      "command": [
        "podman",
        "run",
        "-i",
        "--rm",
        "-e",
        "JIRA_URL",
        "-e",
        "CONFLUENCE_URL",
        "-e",
        "ATLASSIAN_EMAIL",
        "-e",
        "ATLASSIAN_API_TOKEN",
        "ghcr.io/sooperset/mcp-atlassian:latest"
      ],
      "environment": {
        "JIRA_URL": "https://your-domain.atlassian.net",
        "CONFLUENCE_URL": "https://your-domain.atlassian.net/wiki",
        "ATLASSIAN_EMAIL": "{env:ATLASSIAN_EMAIL}",
        "ATLASSIAN_API_TOKEN": "{env:ATLASSIAN_API_TOKEN}"
      },
      "enabled": true
    }
  }
}

Make sure to replace https://your-domain.atlassian.net with your actual Atlassian domain.

Step 4: Launch and Verify OpenCode

Start OpenCode to load the Atlassian MCP tools:

opencode

After launching OpenCode, send a quick initial prompt to the LLM (for example, please reply hello).

To verify that the MCP server is active, press Ctrl+X followed by b to toggle the sidebar/status panel in the OpenCode TUI interface.

OpenCode MCP Sidebar Status

As shown in the sidebar, the atlassian MCP status is visible and connected, indicating that the Jira and Confluence tools are loaded and ready for interaction.

While the MCP server is up and running, you can also verify that the container is active in another terminal tab using podman ps:

podman ps

Output:

CONTAINER ID  IMAGE                                   COMMAND  CREATED        STATUS         PORTS  NAMES
2a8d84b821a8  ghcr.io/sooperset/mcp-atlassian:latest           8 minutes ago  Up 8 minutes         busy_nightingale

You can now use natural language prompts to interact with Jira and Confluence directly:

  • "Summarize Jira ticket PROJ-123 and list its requirements."
  • "Find Confluence pages about deployment workflows and apply those steps."
  • "Update Jira issue PROJ-123 with progress comments."

Troubleshooting & Tips

  • Podman Container Issues: Ensure Podman is installed and able to run rootless containers locally (podman run --rm hello-world).
  • Authentication Failures: Verify that ATLASSIAN_EMAIL and ATLASSIAN_API_TOKEN are properly exported in your environment prior to launching opencode.
  • Domain Formatting: Double-check that JIRA_URL and CONFLUENCE_URL in opencode.json correctly reflect your Atlassian site domain (e.g., https://company.atlassian.net).

HitCount