Testkube Mcp
Let your coding agent trigger Testkube workflows, inspect executions, and pull test artifacts without leaving the editor.
Overview
io.github.kubeshop/testkube-mcp is a MCP server for the Ship phase that lets AI assistants manage Testkube test workflows, executions, and artifacts via the control plane API.
What is this MCP server?
- Manage Testkube test workflows and executions from Claude Code, Cursor, or Codex via stdio MCP
- Authenticate with TK_ACCESS_TOKEN plus org and environment IDs against the Testkube control plane
- Run against docker.io/kubeshop/mcp-server:2.9.1 or configure TK_CONTROL_PLANE_URL and dashboard URLs
- Surface execution status and artifacts for AI-assisted triage after CI or pre-release gates
- Pairs with Kubernetes-native Testkube setups indie teams already use for API and integration suites
- Server schema version 2.9.1
- OCI package docker.io/kubeshop/mcp-server:2.9.1 with stdio transport
- Required secrets: TK_ACCESS_TOKEN; required IDs: TK_ORG_ID, TK_ENV_ID
Community signal: 1.6k GitHub stars.
What problem does it solve?
Jumping between chat, Testkube UI, and CI logs breaks flow when you need quick answers about test runs during a release week.
Who is it for?
Solo builders shipping on Kubernetes who already use Testkube and want agent-driven visibility into workflows and executions.
Skip if: Teams with no Testkube org or who only need local unit tests without a control plane.
What do I get? / Deliverables
After you register the server with your token and env IDs, your agent can query and drive Testkube workflows and artifacts from the same session where you fix code.
- Agent-callable access to Testkube workflow and execution operations
- Artifact and run context surfaced inside the coding session
- Reduced context switching between agent chat and Testkube dashboard
Recommended MCP Servers
Journey fit
How it compares
MCP bridge to Testkube’s test orchestration API, not an agent skill that writes test cases for you.
Common Questions / FAQ
Who is testkube-mcp for?
Indie and small-team developers on Testkube who want Claude Code, Cursor, or Codex to inspect and manage test workflows without leaving the agent.
When should I use testkube-mcp?
Use it during Ship when you gate releases on Testkube executions and need the agent to check status, artifacts, or workflow config on demand.
How do I add testkube-mcp to my agent?
Run the kubeshop MCP server image or stdio package, set TK_ACCESS_TOKEN, TK_ORG_ID, and TK_ENV_ID, then add the server entry in your client’s MCP config.