# Argo Workflows MCP for AI Agents AI Agent Connect > Argo Workflows MCP. Monitor, list, and inspect active pods, crons, and workflow templates directly from any AI agent. It gives you a natural language interface for your Kubernetes orchestrations, letting you debug failed pipeline steps and check scheduled jobs without jumping between tabs or running complex CLI commands. It's the fastest way to get a clear picture of your cluster's health. ## Overview - **Category:** loved-by-devs - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_XgODuKGP8hKPvAVimTmrdsOr5ag915WZ6oPDkNoq/ai-agent-connect - **Tags:** kubernetes, workflow-automation, container-orchestration, ci-cd, infrastructure-as-code, job-scheduling ## Description Stop jumping between your terminal and the Argo Web UI just to figure out why a specific step in your pipeline stalled. This Connector brings your Kubernetes orchestration data directly into your chat interface. You can ask your agent to pull up the status of every running execution or dive deep into a specific resource tree to see exactly which pod hit a snag. It handles the heavy lifting of querying your namespaces, so you can focus on fixing the infrastructure rather than hunting for logs. Whether you're checking on a recurring cron job or auditing your workflow templates, you get the facts instantly. Vinkius makes it simple to plug this into your existing setup, turning your AI into a functional DevOps teammate that understands your specific cluster layout. Instead of manually scrolling through logs or trying to remember the exact namespace for a production job, you can just describe what you're looking for. Your agent can quickly summarize the status of multiple concurrent jobs, identify the specific node that timed out, and even pull up the parameters used in a specific execution. It's about moving from manual inspection to conversational discovery, giving you a high-level overview and a deep-dive capability in a single conversation. ## Tools ### list_workflows List all workflows in a specific Kubernetes namespace. It helps you see what's currently running or pending. ### get_workflow Get the detailed resource tree and status for a single Argo workflow. Use this to find the exact node where a process failed. ### list_archived_workflows List archived workflows from your Argo history. Use this to search for past infrastructure patterns. ### list_workflow_templates List all workflow templates defined in a namespace. This lets you see your reusable components at a glance. ### list_cron_workflows List scheduled cron workflows in a namespace. It's the easiest way to check your recurring jobs. ### get_server_info Get general information about your Argo Workflows instance. Use this to verify your connection and cluster status. ## Prompt Examples **Prompt:** ``` List all active workflows in the 'data-engineering' namespace. ``` **Response:** ``` I found 4 active workflows in 'data-engineering'. Three are executing robustly, but 'etl_nightly_pipeline' has been pending for 22 minutes. Would you like me to get the deep tree view of this pipeline to see which pod is stuck? ``` **Prompt:** ``` What is the detailed status tree of the workflow named 'daily-backup-55x'? ``` **Response:** ``` I've inspected the tree for 'daily-backup-55x'. The master process reports 'Failed'. Specifically, node 's3-upload-step' crashed exiting with code 1. The previous node 'pg-dump' succeeded. It seems to be an S3 permission issue in the pod parameter payload. ``` **Prompt:** ``` Are there any parameterized WorkflowTemplates available for me to run? ``` **Response:** ``` Listing your templates... You have 3 reusable components: `model-training-tmpl` (requires hyperparameters as array input), `data-cleanup` (expects an S3 bucket URI), and `slack-notification-hook`. They are all initialized and stored in the default namespace. ``` ## Capabilities ### Query active workflows See every running or pending execution in a specific namespace. ### Inspect resource trees Get a detailed breakdown of node statuses and pod parameters for any workflow. ### Browse templates View all parameterized WorkflowTemplates available in your cluster. ### View cron jobs List all scheduled recurring tasks across your namespaces. ### Search history Pull up archived workflows to identify historical patterns or past failures. ### Get instance info Retrieve high-level status and information about your Argo Workflows instance. ## Use Cases ### Debugging a stalled ETL A data engineer asks the agent to check why the nightly load is stuck. The agent finds the job and shows the specific node that crashed. ### Auditing production templates A DevOps lead wants to see what reusable components are available. The agent shows all the parameterized options in the production namespace. ### Checking recurring jobs An SRE wants to make sure the daily backups are scheduled. The agent lists all active crons. ### Historical failure analysis A developer wants to know why a job failed last Tuesday. The agent finds the old execution and reports the error. ## Benefits - Faster incident response: Use get_workflow to pinpoint the exact failing pod in a pipeline without hunting through logs. - Easier template management: Quickly browse your list_workflow_templates to see what reusable components you have available. - Better cron oversight: Use list_cron_workflows to see all your scheduled jobs in one list instead of checking a dashboard. - Historical auditing: Use list_archived_workflows to pull up past data and understand why a pipeline behaved a certain way last week. - Reduced context switching: Stop jumping between your IDE, the terminal, and the Argo UI; stay in your chat app. ## How It Works The bottom line is you get a conversational interface for your Kubernetes orchestration without leaving your chat window. 1. Subscribe to the Argo Workflows MCP on Vinkius. 2. Provide your Argo Cluster URL and your RBAC Bearer Token. 3. Ask your AI client to list workflows or inspect specific pods. ## Frequently Asked Questions **What can I do with the Argo Workflows MCP?** You can use it to monitor, list, and inspect your Kubernetes orchestrations. It lets you see what's running, check your scheduled jobs, and find specific pods that might be causing issues. **Can I use this to see why my Kubernetes pipeline failed?** Yes. By asking your agent to inspect a specific workflow, it can pull the resource tree and show you exactly which node failed and why. **How do I check my scheduled cron jobs?** You can simply ask your agent to list your cron workflows. It will pull the list of all scheduled recurring tasks from your namespace. **Can the Argo Workflows MCP see my archived jobs?** Yes, it can pull up archived workflows from your history, which is helpful for auditing past performance or finding old error patterns. **Does this work with any AI client like Claude or Cursor?** Yes, it works with any MCP-compatible client, including Claude, Cursor, and Windsurf, giving you a consistent experience across your tools. **Can I see my workflow templates using this Connector?** Yes. You can ask your agent to list the templates in a namespace to see all your reusable, parameterized components. **How do I connect my Argo cluster to my AI agent?** You just need to provide your Argo Cluster URL and your RBAC Bearer Token in your Connector settings. From there, your agent can query your cluster directly. **Can my AI agent figure out exactly which pod/node failed in an active workflow execution?** Yes. If a workflow fails, you can ask your agent to retrieve the workflow tree by name. The agent uses the `get_workflow` tool to inspect the deeply nested structure, traverse the active nodes, and pinpoint the exact step or container that resulted in an error state without you ever needing to click through the Argo UI. **Can I list only scheduled periodic jobs across my cluster?** Absolutely. You can use the dedicated `list_cron_workflows` capability to isolate and return strictly workloads orchestrated on a time schedule across any namespace, saving you from parsing through thousands of isolated runs. **Do I need to expose my internal Kubernetes API to use this?** No. The integration strictly interfaces with the Argo Server UI/API layer via standard REST traffic using a scoped ServiceAccount Bearer token. Your cluster's overarching master `kube-apiserver` remains safely isolated from external agentic logic.