mcp

This MCP server provides a Tekton-focused Model Context Protocol (MCP) interface. It exposes a set of listing, creation, retrieval, updating, patching and deletion operations for Tekton objects such as Pipelines, PipelineRuns, Tasks, TaskRuns, and StepActions. It also includes Artifact Hub integration to discover and install Tekton Tasks and Pipelines from the community catalog. Use the available operations to query, create from YAML, modify resources, or trigger actions within a Kubernetes cluster that hosts Tekton resources. The server’s toolset is organized into sections for List, Create, Get, Update, Delete, and specialized Artifact Hub operations, enabling programmatic management of Tekton resources through MCP requests.

Key capabilities include:

• List operations with filters like namespace, name prefix, and label selectors (e.g., list_pipelines, list_pipelineruns, list_tasks, list_taskruns, list_stepactions).
• Create operations from YAML definitions (e.g., create_pipeline, create_task, create_pipelinerun, create_taskrun).
• Get operations to retrieve specific resources in JSON or YAML (e.g., get_pipeline, get_task, get_pipelinerun, get_taskrun).
• Update and patch support for existing resources (e.g., update_pipeline, patch_pipeline).
• Delete operations for individual resources (e.g., delete_pipeline, delete_task, delete_pipelinerun, delete_taskrun).
• Start/restart actions for pipelines, tasks, and PipelineRuns to drive Tekton executions (start_pipeline, restart_pipelinerun, etc.).
• Artifact Hub integration to list and install Tekton components from the community catalog (list_artifacthub_tasks, list_artifacthub_pipelines, install_artifacthub_task).

To use these tools, you typically call the MCP server’s endpoints or CLI equivalents using a client that formats requests according to the MCP protocol. Each operation expects specific parameters (namespaced resources, YAML payloads for creation, etc.) and returns structured results that can be consumed by automation scripts, CI pipelines, or other systems integrating Tekton workflows.

Prerequisites:

• Node.js (LTS) and npm installed on your machine
• Network access to install npm packages or access to a registry containing the tektoncd-mcp-server package

Installation steps:

1. Ensure Node.js and npm are installed

node -v
npm -v

2. Install and run the MCP server via npx (as specified by the server configuration)

npm i -g npm@latest           # optional: ensure latest npm
npx -y tektoncd-mcp-server      # or run with specific package if needed

3. Verify the server is running and listening on the expected port. Check the console logs for startup messages. If you need to customize, you can set environment variables as described in additional notes.

4. If you prefer a local install instead of npx, you can install the package locally and run the server script (adjust paths if needed):

npm install tektoncd-mcp-server --save
node node_modules/tektoncd-mcp-server/dist/server.js

5. Configure your MCP client to point at the server endpoint once it is up (default host/port as indicated by the server).

Tips and considerations:

• The MCP server focuses on Tekton-related resources; ensure your Kubernetes cluster has Tekton components installed and accessible from the server environment.
• Use the list and get operations to explore resources before creating or updating them via YAML payloads.
• When creating pipelines or tasks from YAML, validate the YAML against Tekton CRDs prior to submission to avoid retries.
• If you run into authentication or RBAC issues, ensure the server process has the necessary Kubernetes permissions or provide the proper kubeconfig/context.
• Artifact Hub operations can help bootstrap Tekton tasks/pipelines; provide appropriate packageId and optional version when installing.
• Environment variables (if needed) can control endpoints, timeouts, or authentication tokens; expose them via the mcp_config env section as described above.