who_am_i
Current User
Identify the Buildkite user who owns the current API access token. Requires read_user.
Ask your agent in plain language. Each of these maps to a real tool on the server.
who_am_i
Identify the Buildkite user who owns the current API access token. Requires read_user.
list_organizations
Discover accessible Buildkite organizations and their slugs for subsequent tools. Requires read_organizations; organizations without an active plan are omitted.
list_clusters
Discover Buildkite clusters and their UUIDs for create_pipeline. Choose the organization with list_organizations. Requires read_clusters.
list_annotations
Read build annotations and rendered HTML for mutation readback or discovering IDs to delete. Requires read_builds.
get_job_log
DEPRECATED — use `download_job_log` instead. Retrieve the output log for a specific job in a Buildkite build. Useful for debugging failed builds or inspecting command output. Also supports retrieving job environment variables.
get_build
Retrieve detailed information about a specific build including all its jobs/steps. Use this to inspect build results, check individual step statuses, and find job IDs for retrying or inspecting logs.
download_job_log
Prepare a downloadable plain-text job log. Call get_build to discover the job UUID. Optionally inspect its environment variables. Requires read_build_logs and read_job_env when environment is requested.
update_pipeline
Update an existing Buildkite pipeline's settings. Supports changing the name, repository, YAML configuration, branch settings, tags, and visibility. Only provided fields are updated.
manage_job
Retry a failed job or unblock a blocked step in a Buildkite build. Retry re-runs a failed or timed-out job. Unblock releases a manual/block step so the build can continue.
list_builds
List builds across all pipelines or for a specific pipeline. Supports filtering by state, branch, commit, and creation date. Builds are returned newest-first.
get_agent
Inspect a known Buildkite agent, including stopped or disconnected agents. Call list_agents to discover connected agents. Requires read_agents.
list_teams
Discover Buildkite team UUIDs and names for pipeline access assignments. Choose the organization with list_organizations. Requires read_teams.
create_build
Trigger a new build for a Buildkite pipeline. Specify the commit SHA and branch, and optionally set a message, environment variables, and metadata. Use commit "HEAD" to build the latest commit on the branch.
download_artifact
Prepare a downloadable build artifact. Call list_artifacts for artifact and job UUIDs. Requires read_artifacts.
list_artifacts
List artifacts produced by a specific build. Returns file names, paths, sizes, and download URLs. Artifacts are files generated during build execution (test reports, binaries, logs, etc.).
stop_agent
Stop a connected Buildkite agent. By default, the agent finishes its current job before stopping. Use force mode to stop immediately.
list_agents
List connected and stopping Buildkite agents in your organization. Returns names, versions, connection states, metadata tags and current job IDs. Use get_agent to inspect a known stopped or disconnected agent.
manage_build
Request cancellation of a scheduled, running or failing Buildkite build, or rebuild it with its original commit, branch and settings. Cancellation can return an intermediate canceling state; call get_build to check completion. Rebuild creates a new build without fetching a newer branch commit.
delete_annotation
Delete one build annotation identified by list_annotations. Requires write_builds.
create_pipeline
Create a new CI/CD pipeline in your Buildkite organization. Specify a name, repository URL, nonempty YAML configuration and a cluster UUID from list_clusters. Branch settings, team assignments and tags are optional.
create_annotation
Add an annotation to a Buildkite build. Annotations display additional context on the build page and support Markdown/HTML. Useful for attaching test summaries, deployment links, or status information.
archive_pipeline
Archive or unarchive a Buildkite pipeline. Archived pipelines are hidden from the dashboard but retain their data and can be unarchived later.
get_pipeline
Retrieve detailed information about a specific Buildkite pipeline by its slug. Returns configuration, repository, build counts, provider settings, and step definitions.
list_pipelines
List CI/CD pipelines in your Buildkite organization. Returns pipeline names, slugs, repositories, and current build counts. Use this to discover available pipelines before triggering builds or inspecting pipeline details.
delete_pipeline
Permanently delete a Buildkite pipeline. This action is irreversible and will remove all associated builds and data. Consider archiving instead if you may need the data later.
Every client uses the same Magic MCP URL, and so does any other MCP client. Add it once and sign in with SSO.
{"mcpServers": {"metorial": {"url": "<your-magic-mcp-url>"}}}
claude mcp add --transport http metorial <your-magic-mcp-url>
codex mcp add metorial --url <your-magic-mcp-url>codex mcp login metorial
One governed layer for every app your agents reach, not a separate connector per tool.
Discover Buildkite organizations, clusters and teams. Create and manage CI/CD pipelines, trigger and inspect builds, cancel or rebuild builds, retry jobs and unblock manual steps. Inspect agents and stop an agent after its current job. Download job logs and build artifacts, and create, read or delete build annotations.
Common questions about connecting Buildkite to AI agents with Metorial.