Amazon Bedrock Guardrails adds support for coding use cases

This article announces expanded capabilities in Amazon Bedrock Guardrails to protect against harmful content in code-related use cases for generative AI applications.

• Content filters now detect harmful content in code comments, variable names, and string literals
• Prompt leakage detection prevents unintended disclosure of system prompt information
• Denied topics and sensitive information filters safeguard against PII in code structures
• Protects against malicious code injection and intellectual property compromise
• Available across all AWS Regions where Bedrock Guardrails is supported
• Accessible via Amazon Bedrock console and supported APIs

Amazon Bedrock Guardrails now extends security protections to code-based generative AI applications, helping prevent malicious code injection and sensitive data leakage.