This server downloads Jira and Confluence Cloud attachment files to disk using attachment IDs. It's designed to complement the official Atlassian MCP server, which handles search and issue management but doesn't support pulling down actual attachment files. You authenticate with an API token (scoped or classic), then call download_jira_attachment with an ID and output directory. It handles filename sanitization, optional overwrites, and returns a JSON response with file metadata and local path. Requires Python 3.13+ and at least the read:jira:work scope. Confluence attachment support is on the roadmap but not yet implemented.
claude mcp add --transport stdio alyiox-mcp-atlassian-attachments uvx mcp-atlassian-attachments