AIKIDO-2026-557857
braintrust is vulnerable to Insertion of Sensitive Information into Log File
77
High Risk
This Affects:
braintrustTL;DR
The Braintrust-maintained LiteLLM instrumentation records most keyword arguments passed directly to instrumented LiteLLM calls into Braintrust span metadata. Provider credentials and cloud secrets supplied as call arguments, such as api_key, authorization and API-key headers, Azure tokens, AWS access keys, and Vertex AI credentials, are captured and transmitted to Braintrust traces. Users with access to the affected project and its traces can view these stored secrets. The fix records only an allowlist of known-safe LiteLLM request parameters and omits credential fields, headers, transport configuration, and unknown arguments from new span metadata.
Who does this affect?
You are affected if you are using a version that falls within the vulnerable range and you instrument LiteLLM through wrap_litellm(), patch_litellm(), or auto_instrument() while passing provider credentials directly as call arguments.
Background info
braintrust is vulnerable to Insertion of Sensitive Information into Log File in versions 0.2.0 - 0.27.0.
How to fix this
Upgrade the braintrust library to the patch version.
Are You Affected?
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.
I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001Compliant