step-security/Braintrust eval

Automatically run evals on your AI project using Braintrust.

Trust Signals

name	description	required	default
api_key	Your Braintrust API key	yes	—
root	Root directory from which to run evals	no	.
paths	A list of paths (and glob patterns) to include in the evals	no	.
runtime	The runtime to use for evals. Valid values: node, python.	yes	—
package_manager	The package manager to use for evals. Valid values: npm, pnpm, yarn, pip, or uv depending on the runtime.	no	""
use_proxy	Whether to use the Braintrust proxy (to cache LLM calls). Set to 'true' or 'false'.	yes	true
terminate_on_failure	Whether to terminate the evaluation process when an error occurs. Set to 'true' or 'false'.	no	false
github_token	Your GitHub token	yes	${{ github.token }}
step_key	A unique key to identify this step. Do not change this unless you know what you're doing.	yes	${{ github.workflow_ref }}-${{ github.action }}

no outputs