{"slug": "the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer", "title": "The Rundeck MCP Server: AI where your Automation lives by Madeline Zemer", "summary": "PagerDuty announced the general availability of the Runbook Automation / Rundeck MCP Server for Rundeck and self-hosted Runbook Automation customers on version 6.1.0 or higher, letting AI assistants in Claude, Cursor, VS Code, or Codex connect directly to a customer's automation instance. The server lets assistants build jobs, run and report on executions, and write ACL policies under the user's existing permissions, with PagerDuty citing job creation that previously took new users hours now completing in under a minute.", "body_md": "- [PagerDuty](https://www.pagerduty.com/) /\n- [Blog](https://www.pagerduty.com/blog) /\n- [Announcements](https://www.pagerduty.com/blog/category/announcements/) /\n- The Rundeck MCP Server: AI where your Automation lives\n\n# Blog\n\n# The Rundeck MCP Server: AI where your Automation lives\n\n[Madeline Zemer](https://www.pagerduty.com/blog/author/madelinepagerduty-com/)October 7, 2026 | 5 min read\n\n*This blog post is part of PagerDuty’s ongoing series on how we’re helping customers navigate their journey towards autonomous operations. Read on to learn about how PagerDuty’s Runbook Automation / Rundeck MCP Server recently announced in GA builds towards this vision.*\n\nYour team already uses AI assistants in Claude, Cursor, or VS Code. Ask one for help with your Rundeck or Runbook Automation instance and it has to guess. It has no connection to your instance, so it doesn’t know your job schemas, node filter syntax, or ACL policy format. You get a plausible-looking YAML job definition that fails on import, or advice that’s slightly off for your version. Then you paste it into the UI, run it yourself, and report back before the assistant can help with the next step.\n\nThe limitation is access, not model quality. AI in operations is only useful if it can act on the systems where your automation runs, and for a lot of teams that’s Rundeck and Runbook Automation.\n\nWe’ve closed that gap with the [Runbook Automation / Rundeck MCP Server](https://docs.rundeck.com/docs/mcp/), now Generally Available for Rundeck and self-hosted Runbook Automation customers on version 6.1.0 or higher. Add your instance as an MCP connection in Claude, Cursor, VS Code, or Codex, and your assistant can build jobs, run and report on executions, and write ACL policies directly against your instance, under your permissions. Here’s what that looks like across three common tasks.\n\n**Build a working job from a sentence**\n\nOne of the challenges with Runbook Automation has been the ramp. New users need to learn job definitions, plugin configuration, YAML structure, and ACLs before they see a first result. That’s hours of docs and trial and error, and it’s where many teams stall.\n\nWith the MCP Server connected, you describe what you want and the assistant builds it in your instance.\n\nAsk “Do we have any MongoDB jobs already?” and it searches every project and job in seconds. Ask it to “create a job that queries a MongoDB collection using an input I provide at run time, using existing plugins if available,” and it:\n\n- Checks which plugins are installed on your instance and finds the MongoDB Command Runner\n- Builds the job around that plugin with run-time inputs\n- Pulls the database password from Key Storage instead of embedding it\n- Validates the definition against the Rundeck schema before importing, so nothing broken lands on the server\n\nOpen the Rundeck UI and it’s a real job, live and runnable. What used to take a new user hours takes under a minute.\n\n**Run automation and get answers without knowing where anything lives**\n\nDay-to-day operations on an automation instance depend on institutional knowledge: which project a job is in, what it’s called, what its ID is. Operators who don’t carry that map in their heads spend time hunting before they can act, and every question about a run’s outcome means opening the UI and reading logs.\n\nAsk the assistant to “run the API check job” without giving it an exact name or ID. It finds the right job, runs it, waits for it to finish, reads the execution log, and summarizes the result. The same connection lets you pull today’s failed executions, list the jobs in a project, or trigger a recovery job, all in plain language. Naming conventions stop being a prerequisite for getting work done.\n\n**Write and validate ACL policies from intent**\n\nAccess control management can also be a challenge for customers. Automation jobs touch production infrastructure, so limiting who can run what is a security requirement. But ACL policies are YAML with contexts, match rules, and allow and deny lists, and a small mistake silently blocks access or grants too much. It’s the most common point where teams open support tickets or stall.\n\nDescribe the intent instead: “Create an ACL policy that lets the DBA group run jobs in the Databases/MongoDB group of welcome-project, but only view everything else in the project.” The assistant writes the policy, validates the structure before applying it, and explains in plain English what the policy allows. You check the explanation against what you meant, and the syntax is handled for you.\n\n**Bringing it together**\n\nThese three capabilities share one architecture: an MCP connection that gives your assistant direct, permissioned access to your Rundeck or Runbook Automation instance. Because the assistant reads real state (installed plugins, existing jobs, your schema version), it produces artifacts that work on the first import instead of approximations you have to debug. That’s what turns a general-purpose AI assistant into an operator on your platform.\n\nThe safety model is built in. The server uses your own API token, so the assistant can only do what you’re already allowed to do. It validates job definitions and ACL policies before applying them. And it requires human confirmation for destructive actions. You stay in the loop where it matters, and Runbook Automation remains the secured, controlled platform underneath.\n\nFor customers this means faster adoption, with new users productive on day one; faster implementation, with first jobs and access policies taking minutes instead of weeks; and automation that is safe by design.\n\n**Paving the path to autonomous operations**\n\nYour team shouldn’t be hand-translating AI suggestions into YAML and running them yourself before the assistant can help with the next step.\n\nPagerDuty’s approach to autonomous operations:\n\n- Puts intelligent agents to work at scale — handling the noise, accelerating resolution, and keeping you in the loop where it matters\n- Deepens the full incident management lifecycle by empowering teams to resolve incidents faster\n- Broadens the platform and ecosystem with capabilities that help teams prevent incidents from happening\n\nThe Runbook Automation / Rundeck MCP Server moves us closer to that vision. Ready to see how it can put intelligent agents to work on your automation platform for your team? Watch the demo.\n\nLearn more at [https://docs.rundeck.com/docs/mcp/](https://docs.rundeck.com/docs/mcp/) for more information on setup instructions and a best-practices guide.", "url": "https://wpnews.pro/news/the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer", "canonical_source": "https://www.pagerduty.com/blog/announcements/the-rundeck-mcp-server-ai-where-your-automation-lives/", "published_at": "2026-10-07 11:30:42+00:00", "updated_at": "2026-10-07 12:19:01.513833+00:00", "lang": "en", "topics": ["ai-agents", "agent-protocols", "ai-tools", "developer-tools", "mlops"], "entities": ["PagerDuty", "Rundeck", "Runbook Automation", "Rundeck MCP Server", "Madeline Zemer", "Claude", "Cursor", "VS Code"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer", "markdown": "https://wpnews.pro/news/the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer.md", "text": "https://wpnews.pro/news/the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer.txt", "jsonld": "https://wpnews.pro/news/the-rundeck-mcp-server-ai-where-your-automation-lives-by-madeline-zemer.jsonld"}}