working-with-buildkite-builds

Use Buildkite-native MCP tools to fetch build data and logs: call `buildkite:get_build` to retrieve build status, then `get_logs` with the job_uuid to access detailed logs. This gives you direct access to build diagnostics without relying on external GitHub tools, letting you identify failure root causes and reproduce issues locally.

Yes. Fetch build logs, annotations, and status via MCP tools or bundled scripts, then run the same pipeline steps locally. The Skill guides you through interpreting Buildkite states and log data to set up an identical environment, enabling local reproduction and faster iteration on fixes.

Use MCP tools to query build status and logs across repos in a unified workflow. The Skill handles cross-repo pipeline monitoring and post-push workflows, letting you track and diagnose failures consistently without switching between tools or relying on GitHub status pages.

Yes. The Skill applies to PR workflows, post-push monitoring, and cross-repo pipelines. Use MCP tools to fetch build data immediately after pushes or pull requests, access annotations and logs, and diagnose failures before they propagate downstream.

The Skill includes bundled scripts and bktide as fallback mechanisms when MCP tools aren't accessible. Both retrieve build data and logs, ensuring you can still diagnose failures and access annotations even if primary MCP integrations are unavailable.

Buildkite build state is determined by job status, annotations, and log content. The Skill describes functional and technical requirements for accurate data retrieval and state interpretation, helping you avoid misreading build failures caused by annotation placement, log truncation, or transient job errors.