Deployment¶

Deployment Options¶

atlassian-agent supports local stdio, a loopback-only development listener, a least-privilege stdio container, and a remote authenticated HTTPS boundary. Provider endpoint, credential, selector, identity, and trust material are supplied at runtime through AgentConfig; none is stored in this repository.

Installed stdio process¶

{
  "mcpServers": {
    "atlassian": {
      "command": "atlassian-mcp",
      "args": [],
      "env": {"MCP_TOOL_MODE": "intent"}
    }
  }
}

Loopback development listener¶

atlassian-mcp --transport streamable-http --host 127.0.0.1 --port 8000

Do not expose this listener beyond loopback. Network deployments require direct TLS or an explicitly trusted TLS-terminating ingress, configured authentication, exact MCP_ALLOWED_HOSTS, and an exact trusted-proxy CIDR policy.

Least-privilege local container¶

docker run -i --rm \
  --read-only \
  --cap-drop=ALL \
  --security-opt=no-new-privileges \
  --pids-limit=256 \
  --tmpfs /tmp:rw,noexec,nosuid,nodev,size=64m \
  -e TRANSPORT=stdio \
  registry.example.invalid/atlassian-agent@sha256:<digest> atlassian-mcp

The operator projects the selected AgentConfig profile into the process at runtime; the image remains immutable and contains no environment connection profile.

Remote authenticated HTTPS endpoint¶

{
  "mcpServers": {
    "atlassian": {"url": "https://service.example.invalid/mcp"}
  }
}

Store the real remote URL, outbound identity reference, and TLS-profile reference in AgentConfig, not in MCP client JSON or documentation.

This page covers running atlassian-agent as a long-lived service: the transports, a Docker Compose stack, the A2A agent server, putting it behind a Caddy reverse proxy, and giving it a DNS name with Technitium.

atlassian-agent ships two console scripts: an MCP server (atlassian-mcp) exposing a typed, deterministic Atlassian tool surface, and an A2A agent server (atlassian-agent) that drives those tools conversationally. Deploy the MCP server on its own, or deploy both together as a combined stack.

Run the MCP server¶

The transport is selected with --transport (or the TRANSPORT env var):

stdio (default)streamable-httpsse

atlassian-mcp

For IDE / desktop MCP clients that launch the server as a subprocess.

atlassian-mcp --transport streamable-http --host 0.0.0.0 --port 8000

A network server with a /health endpoint and /mcp route.

atlassian-mcp --transport sse --host 0.0.0.0 --port 8000

Health check (HTTP transports):

curl -s http://localhost:8000/health        # {"status":"OK"}

Configuration (environment)¶

atlassian-agent is configured entirely from the environment. The required set for the shared (Cloud) connection:

Var	Default	Meaning
`ATLASSIAN_AGENT_URL`	`http://localhost:8080`	Atlassian base URL (e.g. `https://your-company.atlassian.net`)
`ATLASSIAN_AGENT_USER`	—	Account email / username
`ATLASSIAN_AGENT_TOKEN`	—	API token (Cloud) or password / token (Server)
`ATLASSIAN_TLS_PROFILE`	(unset)	Optional runtime TLS profile selector; verification is mandatory
`HOST`	`0.0.0.0`	Bind address (HTTP transports)
`PORT`	`8000`	Listen port (HTTP transports)
`TRANSPORT`	`stdio`	`stdio`, `streamable-http`, or `sse`

Jira and Confluence Server / Data Center instances may be configured separately with their own credentials (ATLASSIAN_JIRA_SERVER_URL / _USER / _TOKEN / _VERIFY, and ATLASSIAN_CONFLUENCE_SERVER_URL / _USER / _TOKEN / _VERIFY). Each tool group additionally has a *_TOOL toggle (for example JIRA_ISSUE_TOOL, CONFLUENCE_PAGE_TOOL, ATLASSIAN_ADMIN_TOOL) to register only the surface you need. The full set, grouped by product, is documented in .env.example. Copy it to .env and populate only what you use; tools whose credentials are absent remain inactive.

Backing Service¶

Atlassian Jira and Confluence are managed as Atlassian Cloud (a SaaS platform) or as self-operated Server / Data Center products. atlassian-agent is a connector, not a host for those systems, so there is no local backing-platform recipe — only connection configuration is required. Provision an API token from your Atlassian account and point ATLASSIAN_AGENT_URL, ATLASSIAN_AGENT_USER, and ATLASSIAN_AGENT_TOKEN at the instance you intend to manage.

Docker Compose¶

The repository ships docker/mcp.compose.yml. It reads a sibling .env and publishes the HTTP server on :8000:

services:
  atlassian-agent-mcp:
    image: example/atlassian-agent@sha256:<digest>
    container_name: atlassian-agent-mcp
    hostname: atlassian-agent-mcp
    restart: always
    env_file:
      - ../.env
    environment:
      - PYTHONUNBUFFERED=1
      - HOST=0.0.0.0
      - PORT=8000
      - TRANSPORT=streamable-http
    ports:
      - "8000:8000"
    healthcheck:
      test: ["CMD", "python3", "-c", "import urllib.request; urllib.request.urlopen('http://localhost:8000/health')"]
      interval: 30s
      timeout: 10s
      retries: 3

cp .env.example .env          # then edit ATLASSIAN_AGENT_* values
docker compose -f docker/mcp.compose.yml up -d
docker compose -f docker/mcp.compose.yml logs -f

Run the A2A agent server¶

The atlassian-agent console script starts a Pydantic-AI agent that consumes the MCP tool surface and exposes an A2A endpoint (and an optional web UI). Point it at a running MCP server with MCP_URL and select a model provider:

export MCP_URL=http://localhost:8000/mcp
atlassian-agent --provider openai --model-id gpt-4o --api-key sk-...

The repository ships docker/agent.compose.yml, which deploys the MCP server and the agent together. The agent publishes on :9004 and reaches the MCP server by container name:

services:
  atlassian-agent-mcp:
    image: example/atlassian-agent@sha256:<digest>
    hostname: atlassian-agent-mcp
    env_file: [../.env]
    environment:
      - HOST=0.0.0.0
      - PORT=8000
      - TRANSPORT=streamable-http
    ports: ["8000:8000"]

  atlassian-agent-agent:
    image: example/atlassian-agent@sha256:<digest>
    command: ["atlassian-agent"]
    depends_on: [atlassian-agent-mcp]
    env_file: [../.env]
    environment:
      - HOST=0.0.0.0
      - PORT=9004
      - MCP_URL=http://atlassian-agent-mcp:8000/mcp
      - PROVIDER=${PROVIDER:-openai}
      - MODEL_ID=${MODEL_ID:-gpt-4o}
      - ENABLE_WEB_UI=True
    ports: ["9004:9004"]

docker compose -f docker/agent.compose.yml up -d

Behind a Caddy reverse proxy¶

Expose the HTTP server on a hostname with automatic TLS. Add to your Caddyfile:

# Internal (self-signed) — homelab .example.invalid zone
atlassian-agent.example.invalid {
    tls internal
    reverse_proxy atlassian-agent-mcp:8000
}

# Public — automatic Let's Encrypt
atlassian-agent.example.com {
    reverse_proxy atlassian-agent-mcp:8000
}

Reload Caddy:

docker compose -f services/caddy/compose.yml exec caddy caddy reload --config /etc/caddy/Caddyfile

DNS with Technitium¶

Point the hostname at the host running Caddy. Via the Technitium API:

curl -s "http://technitium.example.invalid:5380/api/zones/records/add" \
  --data-urlencode "token=$TECHNITIUM_DNS_TOKEN" \
  --data-urlencode "domain=atlassian-agent.example.invalid" \
  --data-urlencode "zone=arpa" \
  --data-urlencode "type=A" \
  --data-urlencode "ipAddress=192.0.2.10" \
  --data-urlencode "ttl=3600"

…or add an A record atlassian-agent.example.invalid → <caddy-host-ip> in the Technitium web console (http://technitium.example.invalid:5380). The ecosystem technitium-dns-mcp automates this as a tool.

Register with an MCP client¶

Add to your client's mcp_config.json:

{
  "mcpServers": {
    "atlassian-agent": {
      "command": "uv",
      "args": ["run", "atlassian-mcp"],
      "env": {
        "ATLASSIAN_AGENT_URL": "https://your-company.atlassian.net",
        "ATLASSIAN_AGENT_USER": "your-email@example.com",
        "ATLASSIAN_AGENT_TOKEN": "your_api_token"
      }
    }
  }
}

For a remote HTTP server, point the client at http://atlassian-agent.example.invalid/mcp instead.