Connect Buildkite to Codex
Give Codex governed access to Buildkite’s 71 tools through Runlayer, with identity, policy, scanning, and audit on every call.
Set up
Add Buildkite to Codex in four steps
Admins approve the server once. Everyone in Codex then gets it through the same governed connection.
- 1
Approve Buildkite in Runlayer
Add Buildkite from the catalog and choose who can use it and which of its tools are allowed.
- 2
Connect Codex to Runlayer
Runlayer’s CLI adds the Runlayer Plugin to Codex’s local plugin marketplace, and users then enable it in Codex. With direct tool listing turned on for your deployment, Buildkite’s 71 tools show up in Codex’s own tool list (Codex 0.122 or newer, installed through the Runlayer CLI).
Codex setup guide - 3
Sign in to Buildkite
Sign in the way Buildkite requires. Runlayer holds the credentials so they stay out of the client’s configuration.
- 4
Use Buildkite from Codex
Buildkite’s approved tools appear in Codex. Every call is checked against policy and recorded.
Tools
71 Buildkite tools in Codex
Runlayer policy decides which of these Codex users can call.
Access token
Get information about the current API access token including its scopes and UUID
Cancel build
Cancel a running build on a Buildkite pipeline
Compare builds
Compare step outcomes, final-attempt retry counts and execution/scheduling times between builds in one pipeline.
Create annotation
Create an annotation on a build or specific job.
Create build
Trigger a new build on a Buildkite pipeline for a specific commit and branch, with optional environment variables, metadata, and author information
Create cache registry
Create a cache registry in a cluster.
Create cluster
Create a new cluster in an organization
Create cluster queue
Create a new queue in a cluster
Control
What Runlayer can see and stop in Codex
Some controls apply to every client at the gateway. Others depend on what Codex exposes to Runlayer.
Gateway policy (available)
Every Buildkite call from Codex passes through Runlayer, where your policies can allow, alert, mask, or block it.
Audit trail (available)
Each call is recorded with the user, Codex as the client, the tool, and the policy and scanner decisions.
Client hooks (available)
Capture supported prompts, reasoning, local and MCP tool calls, responses, and scanner decisions in real time.
Stop a call before it runs (available)
Runlayer can deny a tool call or permission request before Codex runs it.
Block output after it runs (available)
Runlayer can block a tool’s output before Codex’s model sees it.
Redact output in place (not available)
Codex cannot take masked output, so Runlayer can block a tool’s output but not mask it. The gateway still applies policy to the call.
Frequently asked questions
Connect more
