An open-source platform was built to enable reliable, multi-step AI-driven automation workflows inside Atlassian tools (Jira and Confluence) that survive restarts, retries, and failures. It uses the Model Context Protocol (MCP) to expose Jira and Confluence APIs as typed tools decoupled from agent logic, and Temporal to provide durable workflow execution with checkpointing, retries, and backoff. This approach addresses the limitations of typical short-lived AI agents that fail on long-running or stateful workflows. The platform supports multi-provider LLMs and runs in a Docker Compose stack for local development.
Use Case
Opening the operator briefing
Pulling the full operator breakdown, tooling context, and verification notes.
