AWS bedrock documentation change
Summary
Updated IAM policy example to use wildcard for model ARNs and modified S3 bucket references. Added security note recommending least privilege principle for production environments.
Security assessment
The change adds explicit security guidance about following least privilege principles, but does not address a specific disclosed vulnerability. The wildcard in the example is mitigated by the added security recommendation.
Diff
diff --git a/bedrock/latest/userguide/model-evaluation-type-judge.md b/bedrock/latest/userguide/model-evaluation-type-judge.md index 43076c11b..e779c0b09 100644 --- a/bedrock/latest/userguide/model-evaluation-type-judge.md +++ b/bedrock/latest/userguide/model-evaluation-type-judge.md @@ -23 +23 @@ To create a model evaluation job that uses a LLM as judge, you need access to sp - 2. Create a prompt dataset. Your prompt dataset is a json lines (`jsonl`) formatted dataset that contains the prompts and required ground truth data for the model evaluation job to run successfully. For more information, see [Requirements for custom prompt datasets in model evaluation job that uses a model as judge](./model-evaluation-prompt-datasets-judge.html). + 2. Create a prompt dataset. Your prompt dataset is a json lines (`jsonl`) formatted dataset that contains the prompts and required ground truth data for the model evaluation job to run successfully. For more information, see [Create a custom prompt dataset for a model evaluation job that uses a model as judge](./model-evaluation-prompt-datasets-judge.html). @@ -67,2 +67 @@ The IAM policy must access to both _generator_ and _evaluator_ models. - "arn:aws:bedrock:us-west-2::foundation-model/model-id-of-foundational-model", - "arn:aws:bedrock:us-west-2::foundation-model/model-id-of-foundational-model", + "arn:aws:bedrock:us-west-2::foundation-model/*;", @@ -82,2 +81,2 @@ The IAM policy must access to both _generator_ and _evaluator_ models. - "arn:aws:s3:::my_output_bucket", - "arn:aws:s3:::input_datasets/prompts.jsonl", + "arn:aws:s3:::amzn-s3-demo-destination-bucket/*", + "arn:aws:s3:::input_datasets/prompts.jsonl" @@ -89,0 +89,4 @@ The IAM policy must access to both _generator_ and _evaluator_ models. +###### Note + +This example policy gives permissions for all Amazon Bedrock foundation models. In a production environment, we recomment that you follow the principal of [least privilege](https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html#grant-least-privilege) and only grant permissions for the models you need. +