Mcp Zuul
Debug Zuul CI failures and pipeline state from inside your coding agent instead of juggling the Zuul web UI and raw logs.
Overview
mcp-zuul is a MCP server for the Operate phase that queries Zuul CI for failure analysis, logs, pipeline status, and job config.
What is this MCP server?
- Build failure analysis and log search wired to your Zuul tenant
- Pipeline status and job configuration inspection from MCP tools
- Auth via ZUUL_AUTH_TOKEN, Kerberos (ZUUL_USE_KERBEROS), and configurable ZUUL_URL
- Tool surface trimming with ZUUL_ENABLED_TOOLS / ZUUL_DISABLED_TOOLS
- PyPI mcp-zuul v0.7.0 with stdio, SSE, or streamable-http via MCP_TRANSPORT
- Package version 0.7.0 identifier mcp-zuul on PyPI
- Transports: stdio, sse, streamable-http (MCP_TRANSPORT)
- Four tool domains described: failure analysis, log search, pipeline status, job configuration
Community signal: 6 GitHub stars.
What problem does it solve?
When Zuul blocks your merge, finding the failing job, relevant log lines, and pipeline context across tenants is slow without integrated agent access.
Who is it for?
Solo maintainers on Zuul-backed projects (e.g. OpenStack-style factories) who live in Claude Code or Cursor during incident-style CI debugging.
Skip if: Builders on GitHub Actions-only workflows or anyone without network access and credentials to a Zuul API.
What do I get? / Deliverables
After registration, your agent can pull Zuul pipeline status and searched logs so you fix CI failures and adjust job config with less tab switching.
- Agent-callable Zuul tools for failure analysis and log search
- Live pipeline status reads for your default tenant
- Job configuration visibility for targeted CI fixes
Recommended MCP Servers
Journey fit
CI visibility sits in Operate when pipelines are already running and you need to monitor failures, jobs, and configs in production-like workflows. Monitoring is the canonical shelf because the server exposes build failure analysis, log search, pipeline status, and job configuration against a live Zuul instance.
How it compares
Zuul-specific CI MCP integration, not a generic GitHub Actions skill or local test runner.
Common Questions / FAQ
Who is mcp-zuul for?
Developers who already use Zuul CI and want their AI coding agent to inspect pipelines, logs, and job definitions through MCP.
When should I use mcp-zuul?
Use it when a Zuul pipeline fails post-merge or pre-release and you need log search and failure analysis while iterating in the same session as your code changes.
How do I add mcp-zuul to my agent?
Install mcp-zuul from PyPI via uvx, set ZUUL_URL and ZUUL_DEFAULT_TENANT, add ZUUL_AUTH_TOKEN or Kerberos flags as required, register stdio (or MCP_TRANSPORT http modes), then enable only the tools you need.